├── .gitignore
├── README.md
├── cassandra-cdc
    ├── README.md
    ├── config
    │   ├── cassandra-1-cdc-tmp.yaml
    │   ├── cassandra-2-cdc-tmp.yaml
    │   ├── reader-1.yml
    │   └── reader-2.yml
    ├── docker
    │   ├── Dockerfile
    │   └── cassandra.yaml
    ├── pom.xml
    └── src
    │   └── main
    │       ├── java
    │           └── io
    │           │   └── smartcat
    │           │       └── cassandra
    │           │           └── cdc
    │           │               ├── CustomCommitLogReadHandler.java
    │           │               ├── Reader.java
    │           │               └── YamlUtils.java
    │       └── resources
    │           └── logback.xml
└── cassandra-trigger
    ├── README.md
    ├── pom.xml
    └── src
        └── main
            └── java
                └── io
                    └── smartcat
                        └── cassandra
                            └── trigger
                                └── KafkaTrigger.java


/.gitignore:
--------------------------------------------------------------------------------
1 | **/.classpath
2 | **/.project
3 | **/.settings
4 | **/target
5 | **/logs
6 | **/*.iml
7 | **/*.idea
8 | 


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # cassandra-kafka-connector
2 | cassandra-kafka-connector
3 | 


--------------------------------------------------------------------------------
/cassandra-cdc/README.md:
--------------------------------------------------------------------------------
 1 | # Cassandra Capture Data Change (CDC)
 2 | 
 3 | Purpose of this project is to serve as an example for how to implement read Cassandra CDC.
 4 | 
 5 | ## Create a JAR
 6 | 
 7 | This is a common Maven project with shade plugin to include all dependencies into JAR. To build it, just run:
 8 | 
 9 | `mvn clean install`
10 | 
11 | ## Use JAR
12 | 
13 | In order to start reading CDC commitlogs, run JAR with:
14 | `java -jar cassandra-cdc-0.0.1-SNAPSHOT.jar <path to cdc_raw_directory> <path to configuration>`
15 | 
16 | create trigger in cassandra, JAR file needs to be places under `$CASSANDRA_CONFIG/triggers` directory on every node which will be used as coordinator. Also, path to `KafkaTrigger.yml` (line 37) needs to be adjusted to location where actuall `KafkaTrigger.yml` file is placed. Content of the file should be:
17 | 
18 | ```
19 | bootstrap.servers: cluster_kafka_1:9092,cluster_kafka_2:9092
20 | topic.name: trigger-topic
21 | ```
22 | 
23 | Note that content matches infrastructure setup which is created using `docker-compose` command from `cluster` directory. Docker compose file used is:
24 | 
25 | ```
26 | version: '3.3'
27 | services:
28 |   zookeeper:
29 |     image: wurstmeister/zookeeper:3.4.6
30 |     ports:
31 |       - "2181:2181"
32 |   kafka:
33 |     image: wurstmeister/kafka:0.10.1.1
34 |     ports:
35 |       - 9092
36 |     environment:
37 |       HOSTNAME_COMMAND: "ifconfig | awk '/Bcast:.+/{print $$2}' | awk -F\":\" '{print $$2}'"
38 |       KAFKA_ADVERTISED_PORT: 9092
39 |       KAFKA_ZOOKEEPER_CONNECT: zookeeper:2181
40 |   cassandra-seed:
41 |     image: cassandra-cdc
42 |     ports:
43 |       - 7199
44 |       - 9042
45 |     environment:
46 |       CASSANDRA_CLUSTER_NAME: test-cluster
47 |   cassandra:
48 |     image: cassandra-cdc
49 |     ports:
50 |       - 7199
51 |       - 9042
52 |     environment:
53 |       CASSANDRA_CLUSTER_NAME: test-cluster
54 |       CASSANDRA_SEEDS: cassandra-seed
55 | ```
56 | 
57 | `cassandra-cdc` docker image is custom built for this usage. It includes changes to configuration file, JAR file of reader application and its configuration files.
58 | Whole setup necessary for building `cassandra-cdc` docker image can be found in [docker](docker) directory.
59 | 
60 | ```
61 | FROM cassandra:3.11.0
62 | COPY KafkaTrigger.yml /etc/cassandra/triggers/KafkaTrigger.yml
63 | COPY cassandra-trigger-0.0.1-SNAPSHOT.jar /etc/cassandra/triggers/trigger.jar
64 | CMD ["cassandra", "-f"]
65 | ```
66 | 
67 | If you intend using JAR file in different infrastructure setup (virtual machines, different docker setup, cloud environment) configuration needs to be changed to match that infrastructure.
68 | 
69 | ## Create a trigger
70 | 
71 | To add a trigger to a table, just execute `CREATE TRIGGER kafka_trigger ON movies_by_genre USING 'io.smartcat.cassandra.trigger.KafkaTrigger';`.


--------------------------------------------------------------------------------
/cassandra-cdc/config/cassandra-1-cdc-tmp.yaml:
--------------------------------------------------------------------------------
   1 | # Cassandra storage config YAML
   2 | 
   3 | # NOTE:
   4 | #   See http://wiki.apache.org/cassandra/StorageConfiguration for
   5 | #   full explanations of configuration directives
   6 | # /NOTE
   7 | 
   8 | # The name of the cluster. This is mainly used to prevent machines in
   9 | # one logical cluster from joining another.
  10 | cluster_name: 'Test Cluster'
  11 | 
  12 | # This defines the number of tokens randomly assigned to this node on the ring
  13 | # The more tokens, relative to other nodes, the larger the proportion of data
  14 | # that this node will store. You probably want all nodes to have the same number
  15 | # of tokens assuming they have equal hardware capability.
  16 | #
  17 | # If you leave this unspecified, Cassandra will use the default of 1 token for legacy compatibility,
  18 | # and will use the initial_token as described below.
  19 | #
  20 | # Specifying initial_token will override this setting on the node's initial start,
  21 | # on subsequent starts, this setting will apply even if initial token is set.
  22 | #
  23 | # If you already have a cluster with 1 token per node, and wish to migrate to 
  24 | # multiple tokens per node, see http://wiki.apache.org/cassandra/Operations
  25 | num_tokens: 256
  26 | 
  27 | # Triggers automatic allocation of num_tokens tokens for this node. The allocation
  28 | # algorithm attempts to choose tokens in a way that optimizes replicated load over
  29 | # the nodes in the datacenter for the replication strategy used by the specified
  30 | # keyspace.
  31 | #
  32 | # The load assigned to each node will be close to proportional to its number of
  33 | # vnodes.
  34 | #
  35 | # Only supported with the Murmur3Partitioner.
  36 | # allocate_tokens_for_keyspace: KEYSPACE
  37 | 
  38 | # initial_token allows you to specify tokens manually.  While you can use it with
  39 | # vnodes (num_tokens > 1, above) -- in which case you should provide a 
  40 | # comma-separated list -- it's primarily used when adding nodes to legacy clusters 
  41 | # that do not have vnodes enabled.
  42 | # initial_token:
  43 | 
  44 | # See http://wiki.apache.org/cassandra/HintedHandoff
  45 | # May either be "true" or "false" to enable globally
  46 | hinted_handoff_enabled: true
  47 | 
  48 | # When hinted_handoff_enabled is true, a black list of data centers that will not
  49 | # perform hinted handoff
  50 | # hinted_handoff_disabled_datacenters:
  51 | #    - DC1
  52 | #    - DC2
  53 | 
  54 | # this defines the maximum amount of time a dead host will have hints
  55 | # generated.  After it has been dead this long, new hints for it will not be
  56 | # created until it has been seen alive and gone down again.
  57 | max_hint_window_in_ms: 10800000 # 3 hours
  58 | 
  59 | # Maximum throttle in KBs per second, per delivery thread.  This will be
  60 | # reduced proportionally to the number of nodes in the cluster.  (If there
  61 | # are two nodes in the cluster, each delivery thread will use the maximum
  62 | # rate; if there are three, each will throttle to half of the maximum,
  63 | # since we expect two nodes to be delivering hints simultaneously.)
  64 | hinted_handoff_throttle_in_kb: 1024
  65 | 
  66 | # Number of threads with which to deliver hints;
  67 | # Consider increasing this number when you have multi-dc deployments, since
  68 | # cross-dc handoff tends to be slower
  69 | max_hints_delivery_threads: 2
  70 | 
  71 | # Directory where Cassandra should store hints.
  72 | # If not set, the default directory is $CASSANDRA_HOME/data/hints.
  73 | # hints_directory: /var/lib/cassandra/hints
  74 | hints_directory: /tmp/cdc/cassandra-1/hints
  75 | 
  76 | # How often hints should be flushed from the internal buffers to disk.
  77 | # Will *not* trigger fsync.
  78 | hints_flush_period_in_ms: 10000
  79 | 
  80 | # Maximum size for a single hints file, in megabytes.
  81 | max_hints_file_size_in_mb: 128
  82 | 
  83 | # Compression to apply to the hint files. If omitted, hints files
  84 | # will be written uncompressed. LZ4, Snappy, and Deflate compressors
  85 | # are supported.
  86 | #hints_compression:
  87 | #   - class_name: LZ4Compressor
  88 | #     parameters:
  89 | #         -
  90 | 
  91 | # Maximum throttle in KBs per second, total. This will be
  92 | # reduced proportionally to the number of nodes in the cluster.
  93 | batchlog_replay_throttle_in_kb: 1024
  94 | 
  95 | # Authentication backend, implementing IAuthenticator; used to identify users
  96 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthenticator,
  97 | # PasswordAuthenticator}.
  98 | #
  99 | # - AllowAllAuthenticator performs no checks - set it to disable authentication.
 100 | # - PasswordAuthenticator relies on username/password pairs to authenticate
 101 | #   users. It keeps usernames and hashed passwords in system_auth.roles table.
 102 | #   Please increase system_auth keyspace replication factor if you use this authenticator.
 103 | #   If using PasswordAuthenticator, CassandraRoleManager must also be used (see below)
 104 | authenticator: AllowAllAuthenticator
 105 | 
 106 | # Authorization backend, implementing IAuthorizer; used to limit access/provide permissions
 107 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthorizer,
 108 | # CassandraAuthorizer}.
 109 | #
 110 | # - AllowAllAuthorizer allows any action to any user - set it to disable authorization.
 111 | # - CassandraAuthorizer stores permissions in system_auth.role_permissions table. Please
 112 | #   increase system_auth keyspace replication factor if you use this authorizer.
 113 | authorizer: AllowAllAuthorizer
 114 | 
 115 | # Part of the Authentication & Authorization backend, implementing IRoleManager; used
 116 | # to maintain grants and memberships between roles.
 117 | # Out of the box, Cassandra provides org.apache.cassandra.auth.CassandraRoleManager,
 118 | # which stores role information in the system_auth keyspace. Most functions of the
 119 | # IRoleManager require an authenticated login, so unless the configured IAuthenticator
 120 | # actually implements authentication, most of this functionality will be unavailable.
 121 | #
 122 | # - CassandraRoleManager stores role data in the system_auth keyspace. Please
 123 | #   increase system_auth keyspace replication factor if you use this role manager.
 124 | role_manager: CassandraRoleManager
 125 | 
 126 | # Validity period for roles cache (fetching granted roles can be an expensive
 127 | # operation depending on the role manager, CassandraRoleManager is one example)
 128 | # Granted roles are cached for authenticated sessions in AuthenticatedUser and
 129 | # after the period specified here, become eligible for (async) reload.
 130 | # Defaults to 2000, set to 0 to disable caching entirely.
 131 | # Will be disabled automatically for AllowAllAuthenticator.
 132 | roles_validity_in_ms: 2000
 133 | 
 134 | # Refresh interval for roles cache (if enabled).
 135 | # After this interval, cache entries become eligible for refresh. Upon next
 136 | # access, an async reload is scheduled and the old value returned until it
 137 | # completes. If roles_validity_in_ms is non-zero, then this must be
 138 | # also.
 139 | # Defaults to the same value as roles_validity_in_ms.
 140 | # roles_update_interval_in_ms: 2000
 141 | 
 142 | # Validity period for permissions cache (fetching permissions can be an
 143 | # expensive operation depending on the authorizer, CassandraAuthorizer is
 144 | # one example). Defaults to 2000, set to 0 to disable.
 145 | # Will be disabled automatically for AllowAllAuthorizer.
 146 | permissions_validity_in_ms: 2000
 147 | 
 148 | # Refresh interval for permissions cache (if enabled).
 149 | # After this interval, cache entries become eligible for refresh. Upon next
 150 | # access, an async reload is scheduled and the old value returned until it
 151 | # completes. If permissions_validity_in_ms is non-zero, then this must be
 152 | # also.
 153 | # Defaults to the same value as permissions_validity_in_ms.
 154 | # permissions_update_interval_in_ms: 2000
 155 | 
 156 | # Validity period for credentials cache. This cache is tightly coupled to
 157 | # the provided PasswordAuthenticator implementation of IAuthenticator. If
 158 | # another IAuthenticator implementation is configured, this cache will not
 159 | # be automatically used and so the following settings will have no effect.
 160 | # Please note, credentials are cached in their encrypted form, so while
 161 | # activating this cache may reduce the number of queries made to the
 162 | # underlying table, it may not  bring a significant reduction in the
 163 | # latency of individual authentication attempts.
 164 | # Defaults to 2000, set to 0 to disable credentials caching.
 165 | credentials_validity_in_ms: 2000
 166 | 
 167 | # Refresh interval for credentials cache (if enabled).
 168 | # After this interval, cache entries become eligible for refresh. Upon next
 169 | # access, an async reload is scheduled and the old value returned until it
 170 | # completes. If credentials_validity_in_ms is non-zero, then this must be
 171 | # also.
 172 | # Defaults to the same value as credentials_validity_in_ms.
 173 | # credentials_update_interval_in_ms: 2000
 174 | 
 175 | # The partitioner is responsible for distributing groups of rows (by
 176 | # partition key) across nodes in the cluster.  You should leave this
 177 | # alone for new clusters.  The partitioner can NOT be changed without
 178 | # reloading all data, so when upgrading you should set this to the
 179 | # same partitioner you were already using.
 180 | #
 181 | # Besides Murmur3Partitioner, partitioners included for backwards
 182 | # compatibility include RandomPartitioner, ByteOrderedPartitioner, and
 183 | # OrderPreservingPartitioner.
 184 | #
 185 | partitioner: org.apache.cassandra.dht.Murmur3Partitioner
 186 | 
 187 | # Directories where Cassandra should store data on disk.  Cassandra
 188 | # will spread data evenly across them, subject to the granularity of
 189 | # the configured compaction strategy.
 190 | # If not set, the default directory is $CASSANDRA_HOME/data/data.
 191 | # data_file_directories:
 192 | #     - /var/lib/cassandra/data
 193 | data_file_directories:
 194 |   - /tmp/cdc/cassandra-1/data
 195 | 
 196 | # commit log.  when running on magnetic HDD, this should be a
 197 | # separate spindle than the data directories.
 198 | # If not set, the default directory is $CASSANDRA_HOME/data/commitlog.
 199 | # commitlog_directory: /var/lib/cassandra/commitlog
 200 | commitlog_directory: /tmp/cdc/cassandra-1/commitlog
 201 | 
 202 | # Enable / disable CDC functionality on a per-node basis. This modifies the logic used
 203 | # for write path allocation rejection (standard: never reject. cdc: reject Mutation
 204 | # containing a CDC-enabled table if at space limit in cdc_raw_directory).
 205 | cdc_enabled: true
 206 | 
 207 | # CommitLogSegments are moved to this directory on flush if cdc_enabled: true and the
 208 | # segment contains mutations for a CDC-enabled table. This should be placed on a
 209 | # separate spindle than the data directories. If not set, the default directory is
 210 | # $CASSANDRA_HOME/data/cdc_raw.
 211 | # cdc_raw_directory: /var/lib/cassandra/cdc_raw
 212 | cdc_raw_directory: /tmp/cdc/cassandra-1/cdc_raw
 213 | 
 214 | # Policy for data disk failures:
 215 | #
 216 | # die
 217 | #   shut down gossip and client transports and kill the JVM for any fs errors or
 218 | #   single-sstable errors, so the node can be replaced.
 219 | #
 220 | # stop_paranoid
 221 | #   shut down gossip and client transports even for single-sstable errors,
 222 | #   kill the JVM for errors during startup.
 223 | #
 224 | # stop
 225 | #   shut down gossip and client transports, leaving the node effectively dead, but
 226 | #   can still be inspected via JMX, kill the JVM for errors during startup.
 227 | #
 228 | # best_effort
 229 | #    stop using the failed disk and respond to requests based on
 230 | #    remaining available sstables.  This means you WILL see obsolete
 231 | #    data at CL.ONE!
 232 | #
 233 | # ignore
 234 | #    ignore fatal errors and let requests fail, as in pre-1.2 Cassandra
 235 | disk_failure_policy: stop
 236 | 
 237 | # Policy for commit disk failures:
 238 | #
 239 | # die
 240 | #   shut down gossip and Thrift and kill the JVM, so the node can be replaced.
 241 | #
 242 | # stop
 243 | #   shut down gossip and Thrift, leaving the node effectively dead, but
 244 | #   can still be inspected via JMX.
 245 | #
 246 | # stop_commit
 247 | #   shutdown the commit log, letting writes collect but
 248 | #   continuing to service reads, as in pre-2.0.5 Cassandra
 249 | #
 250 | # ignore
 251 | #   ignore fatal errors and let the batches fail
 252 | commit_failure_policy: stop
 253 | 
 254 | # Maximum size of the native protocol prepared statement cache
 255 | #
 256 | # Valid values are either "auto" (omitting the value) or a value greater 0.
 257 | #
 258 | # Note that specifying a too large value will result in long running GCs and possbily
 259 | # out-of-memory errors. Keep the value at a small fraction of the heap.
 260 | #
 261 | # If you constantly see "prepared statements discarded in the last minute because
 262 | # cache limit reached" messages, the first step is to investigate the root cause
 263 | # of these messages and check whether prepared statements are used correctly -
 264 | # i.e. use bind markers for variable parts.
 265 | #
 266 | # Do only change the default value, if you really have more prepared statements than
 267 | # fit in the cache. In most cases it is not neccessary to change this value.
 268 | # Constantly re-preparing statements is a performance penalty.
 269 | #
 270 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 271 | prepared_statements_cache_size_mb:
 272 | 
 273 | # Maximum size of the Thrift prepared statement cache
 274 | #
 275 | # If you do not use Thrift at all, it is safe to leave this value at "auto".
 276 | #
 277 | # See description of 'prepared_statements_cache_size_mb' above for more information.
 278 | #
 279 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 280 | thrift_prepared_statements_cache_size_mb:
 281 | 
 282 | # Maximum size of the key cache in memory.
 283 | #
 284 | # Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at the
 285 | # minimum, sometimes more. The key cache is fairly tiny for the amount of
 286 | # time it saves, so it's worthwhile to use it at large numbers.
 287 | # The row cache saves even more time, but must contain the entire row,
 288 | # so it is extremely space-intensive. It's best to only use the
 289 | # row cache if you have hot rows or static rows.
 290 | #
 291 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 292 | #
 293 | # Default value is empty to make it "auto" (min(5% of Heap (in MB), 100MB)). Set to 0 to disable key cache.
 294 | key_cache_size_in_mb:
 295 | 
 296 | # Duration in seconds after which Cassandra should
 297 | # save the key cache. Caches are saved to saved_caches_directory as
 298 | # specified in this configuration file.
 299 | #
 300 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 301 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 302 | # has limited use.
 303 | #
 304 | # Default is 14400 or 4 hours.
 305 | key_cache_save_period: 14400
 306 | 
 307 | # Number of keys from the key cache to save
 308 | # Disabled by default, meaning all keys are going to be saved
 309 | # key_cache_keys_to_save: 100
 310 | 
 311 | # Row cache implementation class name. Available implementations:
 312 | #
 313 | # org.apache.cassandra.cache.OHCProvider
 314 | #   Fully off-heap row cache implementation (default).
 315 | #
 316 | # org.apache.cassandra.cache.SerializingCacheProvider
 317 | #   This is the row cache implementation availabile
 318 | #   in previous releases of Cassandra.
 319 | # row_cache_class_name: org.apache.cassandra.cache.OHCProvider
 320 | 
 321 | # Maximum size of the row cache in memory.
 322 | # Please note that OHC cache implementation requires some additional off-heap memory to manage
 323 | # the map structures and some in-flight memory during operations before/after cache entries can be
 324 | # accounted against the cache capacity. This overhead is usually small compared to the whole capacity.
 325 | # Do not specify more memory that the system can afford in the worst usual situation and leave some
 326 | # headroom for OS block level cache. Do never allow your system to swap.
 327 | #
 328 | # Default value is 0, to disable row caching.
 329 | row_cache_size_in_mb: 0
 330 | 
 331 | # Duration in seconds after which Cassandra should save the row cache.
 332 | # Caches are saved to saved_caches_directory as specified in this configuration file.
 333 | #
 334 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 335 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 336 | # has limited use.
 337 | #
 338 | # Default is 0 to disable saving the row cache.
 339 | row_cache_save_period: 0
 340 | 
 341 | # Number of keys from the row cache to save.
 342 | # Specify 0 (which is the default), meaning all keys are going to be saved
 343 | # row_cache_keys_to_save: 100
 344 | 
 345 | # Maximum size of the counter cache in memory.
 346 | #
 347 | # Counter cache helps to reduce counter locks' contention for hot counter cells.
 348 | # In case of RF = 1 a counter cache hit will cause Cassandra to skip the read before
 349 | # write entirely. With RF > 1 a counter cache hit will still help to reduce the duration
 350 | # of the lock hold, helping with hot counter cell updates, but will not allow skipping
 351 | # the read entirely. Only the local (clock, count) tuple of a counter cell is kept
 352 | # in memory, not the whole counter, so it's relatively cheap.
 353 | #
 354 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 355 | #
 356 | # Default value is empty to make it "auto" (min(2.5% of Heap (in MB), 50MB)). Set to 0 to disable counter cache.
 357 | # NOTE: if you perform counter deletes and rely on low gcgs, you should disable the counter cache.
 358 | counter_cache_size_in_mb:
 359 | 
 360 | # Duration in seconds after which Cassandra should
 361 | # save the counter cache (keys only). Caches are saved to saved_caches_directory as
 362 | # specified in this configuration file.
 363 | #
 364 | # Default is 7200 or 2 hours.
 365 | counter_cache_save_period: 7200
 366 | 
 367 | # Number of keys from the counter cache to save
 368 | # Disabled by default, meaning all keys are going to be saved
 369 | # counter_cache_keys_to_save: 100
 370 | 
 371 | # saved caches
 372 | # If not set, the default directory is $CASSANDRA_HOME/data/saved_caches.
 373 | # saved_caches_directory: /var/lib/cassandra/saved_caches
 374 | 
 375 | # commitlog_sync may be either "periodic" or "batch." 
 376 | # 
 377 | # When in batch mode, Cassandra won't ack writes until the commit log
 378 | # has been fsynced to disk.  It will wait
 379 | # commitlog_sync_batch_window_in_ms milliseconds between fsyncs.
 380 | # This window should be kept short because the writer threads will
 381 | # be unable to do extra work while waiting.  (You may need to increase
 382 | # concurrent_writes for the same reason.)
 383 | #
 384 | # commitlog_sync: batch
 385 | # commitlog_sync_batch_window_in_ms: 2
 386 | #
 387 | # the other option is "periodic" where writes may be acked immediately
 388 | # and the CommitLog is simply synced every commitlog_sync_period_in_ms
 389 | # milliseconds. 
 390 | commitlog_sync: periodic
 391 | commitlog_sync_period_in_ms: 10000
 392 | 
 393 | # The size of the individual commitlog file segments.  A commitlog
 394 | # segment may be archived, deleted, or recycled once all the data
 395 | # in it (potentially from each columnfamily in the system) has been
 396 | # flushed to sstables.
 397 | #
 398 | # The default size is 32, which is almost always fine, but if you are
 399 | # archiving commitlog segments (see commitlog_archiving.properties),
 400 | # then you probably want a finer granularity of archiving; 8 or 16 MB
 401 | # is reasonable.
 402 | # Max mutation size is also configurable via max_mutation_size_in_kb setting in
 403 | # cassandra.yaml. The default is half the size commitlog_segment_size_in_mb * 1024.
 404 | #
 405 | # NOTE: If max_mutation_size_in_kb is set explicitly then commitlog_segment_size_in_mb must
 406 | # be set to at least twice the size of max_mutation_size_in_kb / 1024
 407 | #
 408 | commitlog_segment_size_in_mb: 1
 409 | 
 410 | # Compression to apply to the commit log. If omitted, the commit log
 411 | # will be written uncompressed.  LZ4, Snappy, and Deflate compressors
 412 | # are supported.
 413 | # commitlog_compression:
 414 | #   - class_name: LZ4Compressor
 415 | #     parameters:
 416 | #         -
 417 | 
 418 | # any class that implements the SeedProvider interface and has a
 419 | # constructor that takes a Map<String, String> of parameters will do.
 420 | seed_provider:
 421 |     # Addresses of hosts that are deemed contact points. 
 422 |     # Cassandra nodes use this list of hosts to find each other and learn
 423 |     # the topology of the ring.  You must change this if you are running
 424 |     # multiple nodes!
 425 |     - class_name: org.apache.cassandra.locator.SimpleSeedProvider
 426 |       parameters:
 427 |           # seeds is actually a comma-delimited list of addresses.
 428 |           # Ex: "<ip1>,<ip2>,<ip3>"
 429 |           - seeds: "127.0.0.1"
 430 | 
 431 | # For workloads with more data than can fit in memory, Cassandra's
 432 | # bottleneck will be reads that need to fetch data from
 433 | # disk. "concurrent_reads" should be set to (16 * number_of_drives) in
 434 | # order to allow the operations to enqueue low enough in the stack
 435 | # that the OS and drives can reorder them. Same applies to
 436 | # "concurrent_counter_writes", since counter writes read the current
 437 | # values before incrementing and writing them back.
 438 | #
 439 | # On the other hand, since writes are almost never IO bound, the ideal
 440 | # number of "concurrent_writes" is dependent on the number of cores in
 441 | # your system; (8 * number_of_cores) is a good rule of thumb.
 442 | concurrent_reads: 32
 443 | concurrent_writes: 32
 444 | concurrent_counter_writes: 32
 445 | 
 446 | # For materialized view writes, as there is a read involved, so this should
 447 | # be limited by the less of concurrent reads or concurrent writes.
 448 | concurrent_materialized_view_writes: 32
 449 | 
 450 | # Maximum memory to use for sstable chunk cache and buffer pooling.
 451 | # 32MB of this are reserved for pooling buffers, the rest is used as an
 452 | # cache that holds uncompressed sstable chunks.
 453 | # Defaults to the smaller of 1/4 of heap or 512MB. This pool is allocated off-heap,
 454 | # so is in addition to the memory allocated for heap. The cache also has on-heap
 455 | # overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size
 456 | # if the default 64k chunk size is used).
 457 | # Memory is only allocated when needed.
 458 | # file_cache_size_in_mb: 512
 459 | 
 460 | # Flag indicating whether to allocate on or off heap when the sstable buffer
 461 | # pool is exhausted, that is when it has exceeded the maximum memory
 462 | # file_cache_size_in_mb, beyond which it will not cache buffers but allocate on request.
 463 | 
 464 | # buffer_pool_use_heap_if_exhausted: true
 465 | 
 466 | # The strategy for optimizing disk read
 467 | # Possible values are:
 468 | # ssd (for solid state disks, the default)
 469 | # spinning (for spinning disks)
 470 | # disk_optimization_strategy: ssd
 471 | 
 472 | # Total permitted memory to use for memtables. Cassandra will stop
 473 | # accepting writes when the limit is exceeded until a flush completes,
 474 | # and will trigger a flush based on memtable_cleanup_threshold
 475 | # If omitted, Cassandra will set both to 1/4 the size of the heap.
 476 | # memtable_heap_space_in_mb: 2048
 477 | # memtable_offheap_space_in_mb: 2048
 478 | 
 479 | # memtable_cleanup_threshold is deprecated. The default calculation
 480 | # is the only reasonable choice. See the comments on  memtable_flush_writers
 481 | # for more information.
 482 | #
 483 | # Ratio of occupied non-flushing memtable size to total permitted size
 484 | # that will trigger a flush of the largest memtable. Larger mct will
 485 | # mean larger flushes and hence less compaction, but also less concurrent
 486 | # flush activity which can make it difficult to keep your disks fed
 487 | # under heavy write load.
 488 | #
 489 | # memtable_cleanup_threshold defaults to 1 / (memtable_flush_writers + 1)
 490 | # memtable_cleanup_threshold: 0.11
 491 | 
 492 | # Specify the way Cassandra allocates and manages memtable memory.
 493 | # Options are:
 494 | #
 495 | # heap_buffers
 496 | #   on heap nio buffers
 497 | #
 498 | # offheap_buffers
 499 | #   off heap (direct) nio buffers
 500 | #
 501 | # offheap_objects
 502 | #    off heap objects
 503 | memtable_allocation_type: heap_buffers
 504 | 
 505 | # Total space to use for commit logs on disk.
 506 | #
 507 | # If space gets above this value, Cassandra will flush every dirty CF
 508 | # in the oldest segment and remove it.  So a small total commitlog space
 509 | # will tend to cause more flush activity on less-active columnfamilies.
 510 | #
 511 | # The default value is the smaller of 8192, and 1/4 of the total space
 512 | # of the commitlog volume.
 513 | #
 514 | # commitlog_total_space_in_mb: 8192
 515 | commitlog_total_space_in_mb: 16
 516 | 
 517 | # This sets the number of memtable flush writer threads per disk
 518 | # as well as the total number of memtables that can be flushed concurrently.
 519 | # These are generally a combination of compute and IO bound.
 520 | #
 521 | # Memtable flushing is more CPU efficient than memtable ingest and a single thread
 522 | # can keep up with the ingest rate of a whole server on a single fast disk
 523 | # until it temporarily becomes IO bound under contention typically with compaction.
 524 | # At that point you need multiple flush threads. At some point in the future
 525 | # it may become CPU bound all the time.
 526 | #
 527 | # You can tell if flushing is falling behind using the MemtablePool.BlockedOnAllocation
 528 | # metric which should be 0, but will be non-zero if threads are blocked waiting on flushing
 529 | # to free memory.
 530 | #
 531 | # memtable_flush_writers defaults to two for a single data directory.
 532 | # This means that two  memtables can be flushed concurrently to the single data directory.
 533 | # If you have multiple data directories the default is one memtable flushing at a time
 534 | # but the flush will use a thread per data directory so you will get two or more writers.
 535 | #
 536 | # Two is generally enough to flush on a fast disk [array] mounted as a single data directory.
 537 | # Adding more flush writers will result in smaller more frequent flushes that introduce more
 538 | # compaction overhead.
 539 | #
 540 | # There is a direct tradeoff between number of memtables that can be flushed concurrently
 541 | # and flush size and frequency. More is not better you just need enough flush writers
 542 | # to never stall waiting for flushing to free memory.
 543 | #
 544 | #memtable_flush_writers: 2
 545 | 
 546 | # Total space to use for change-data-capture logs on disk.
 547 | #
 548 | # If space gets above this value, Cassandra will throw WriteTimeoutException
 549 | # on Mutations including tables with CDC enabled. A CDCCompactor is responsible
 550 | # for parsing the raw CDC logs and deleting them when parsing is completed.
 551 | #
 552 | # The default value is the min of 4096 mb and 1/8th of the total space
 553 | # of the drive where cdc_raw_directory resides.
 554 | # cdc_total_space_in_mb: 4096
 555 | cdc_total_space_in_mb: 4096
 556 | 
 557 | # When we hit our cdc_raw limit and the CDCCompactor is either running behind
 558 | # or experiencing backpressure, we check at the following interval to see if any
 559 | # new space for cdc-tracked tables has been made available. Default to 250ms
 560 | # cdc_free_space_check_interval_ms: 250
 561 | cdc_free_space_check_interval_ms: 250
 562 | 
 563 | # A fixed memory pool size in MB for for SSTable index summaries. If left
 564 | # empty, this will default to 5% of the heap size. If the memory usage of
 565 | # all index summaries exceeds this limit, SSTables with low read rates will
 566 | # shrink their index summaries in order to meet this limit.  However, this
 567 | # is a best-effort process. In extreme conditions Cassandra may need to use
 568 | # more than this amount of memory.
 569 | index_summary_capacity_in_mb:
 570 | 
 571 | # How frequently index summaries should be resampled.  This is done
 572 | # periodically to redistribute memory from the fixed-size pool to sstables
 573 | # proportional their recent read rates.  Setting to -1 will disable this
 574 | # process, leaving existing index summaries at their current sampling level.
 575 | index_summary_resize_interval_in_minutes: 60
 576 | 
 577 | # Whether to, when doing sequential writing, fsync() at intervals in
 578 | # order to force the operating system to flush the dirty
 579 | # buffers. Enable this to avoid sudden dirty buffer flushing from
 580 | # impacting read latencies. Almost always a good idea on SSDs; not
 581 | # necessarily on platters.
 582 | trickle_fsync: false
 583 | trickle_fsync_interval_in_kb: 10240
 584 | 
 585 | # TCP port, for commands and data
 586 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 587 | storage_port: 7000
 588 | 
 589 | # SSL port, for encrypted communication.  Unused unless enabled in
 590 | # encryption_options
 591 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 592 | ssl_storage_port: 7001
 593 | 
 594 | # Address or interface to bind to and tell other Cassandra nodes to connect to.
 595 | # You _must_ change this if you want multiple nodes to be able to communicate!
 596 | #
 597 | # Set listen_address OR listen_interface, not both.
 598 | #
 599 | # Leaving it blank leaves it up to InetAddress.getLocalHost(). This
 600 | # will always do the Right Thing _if_ the node is properly configured
 601 | # (hostname, name resolution, etc), and the Right Thing is to use the
 602 | # address associated with the hostname (it might not be).
 603 | #
 604 | # Setting listen_address to 0.0.0.0 is always wrong.
 605 | #
 606 | listen_address: localhost
 607 | 
 608 | # Set listen_address OR listen_interface, not both. Interfaces must correspond
 609 | # to a single address, IP aliasing is not supported.
 610 | # listen_interface: eth0
 611 | 
 612 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 613 | # you can specify which should be chosen using listen_interface_prefer_ipv6. If false the first ipv4
 614 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 615 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 616 | # listen_interface_prefer_ipv6: false
 617 | 
 618 | # Address to broadcast to other Cassandra nodes
 619 | # Leaving this blank will set it to the same value as listen_address
 620 | # broadcast_address: 1.2.3.4
 621 | 
 622 | # When using multiple physical network interfaces, set this
 623 | # to true to listen on broadcast_address in addition to
 624 | # the listen_address, allowing nodes to communicate in both
 625 | # interfaces.
 626 | # Ignore this property if the network configuration automatically
 627 | # routes  between the public and private networks such as EC2.
 628 | # listen_on_broadcast_address: false
 629 | 
 630 | # Internode authentication backend, implementing IInternodeAuthenticator;
 631 | # used to allow/disallow connections from peer nodes.
 632 | # internode_authenticator: org.apache.cassandra.auth.AllowAllInternodeAuthenticator
 633 | 
 634 | # Whether to start the native transport server.
 635 | # Please note that the address on which the native transport is bound is the
 636 | # same as the rpc_address. The port however is different and specified below.
 637 | start_native_transport: true
 638 | # port for the CQL native transport to listen for clients on
 639 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 640 | native_transport_port: 9042
 641 | # Enabling native transport encryption in client_encryption_options allows you to either use
 642 | # encryption for the standard port or to use a dedicated, additional port along with the unencrypted
 643 | # standard native_transport_port.
 644 | # Enabling client encryption and keeping native_transport_port_ssl disabled will use encryption
 645 | # for native_transport_port. Setting native_transport_port_ssl to a different value
 646 | # from native_transport_port will use encryption for native_transport_port_ssl while
 647 | # keeping native_transport_port unencrypted.
 648 | # native_transport_port_ssl: 9142
 649 | # The maximum threads for handling requests when the native transport is used.
 650 | # This is similar to rpc_max_threads though the default differs slightly (and
 651 | # there is no native_transport_min_threads, idle threads will always be stopped
 652 | # after 30 seconds).
 653 | # native_transport_max_threads: 128
 654 | #
 655 | # The maximum size of allowed frame. Frame (requests) larger than this will
 656 | # be rejected as invalid. The default is 256MB. If you're changing this parameter,
 657 | # you may want to adjust max_value_size_in_mb accordingly.
 658 | # native_transport_max_frame_size_in_mb: 256
 659 | 
 660 | # The maximum number of concurrent client connections.
 661 | # The default is -1, which means unlimited.
 662 | # native_transport_max_concurrent_connections: -1
 663 | 
 664 | # The maximum number of concurrent client connections per source ip.
 665 | # The default is -1, which means unlimited.
 666 | # native_transport_max_concurrent_connections_per_ip: -1
 667 | 
 668 | # Whether to start the thrift rpc server.
 669 | start_rpc: false
 670 | 
 671 | # The address or interface to bind the Thrift RPC service and native transport
 672 | # server to.
 673 | #
 674 | # Set rpc_address OR rpc_interface, not both.
 675 | #
 676 | # Leaving rpc_address blank has the same effect as on listen_address
 677 | # (i.e. it will be based on the configured hostname of the node).
 678 | #
 679 | # Note that unlike listen_address, you can specify 0.0.0.0, but you must also
 680 | # set broadcast_rpc_address to a value other than 0.0.0.0.
 681 | #
 682 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 683 | rpc_address: localhost
 684 | 
 685 | # Set rpc_address OR rpc_interface, not both. Interfaces must correspond
 686 | # to a single address, IP aliasing is not supported.
 687 | # rpc_interface: eth1
 688 | 
 689 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 690 | # you can specify which should be chosen using rpc_interface_prefer_ipv6. If false the first ipv4
 691 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 692 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 693 | # rpc_interface_prefer_ipv6: false
 694 | 
 695 | # port for Thrift to listen for clients on
 696 | rpc_port: 9160
 697 | 
 698 | # RPC address to broadcast to drivers and other Cassandra nodes. This cannot
 699 | # be set to 0.0.0.0. If left blank, this will be set to the value of
 700 | # rpc_address. If rpc_address is set to 0.0.0.0, broadcast_rpc_address must
 701 | # be set.
 702 | # broadcast_rpc_address: 1.2.3.4
 703 | 
 704 | # enable or disable keepalive on rpc/native connections
 705 | rpc_keepalive: true
 706 | 
 707 | # Cassandra provides two out-of-the-box options for the RPC Server:
 708 | #
 709 | # sync
 710 | #   One thread per thrift connection. For a very large number of clients, memory
 711 | #   will be your limiting factor. On a 64 bit JVM, 180KB is the minimum stack size
 712 | #   per thread, and that will correspond to your use of virtual memory (but physical memory
 713 | #   may be limited depending on use of stack space).
 714 | #
 715 | # hsha
 716 | #   Stands for "half synchronous, half asynchronous." All thrift clients are handled
 717 | #   asynchronously using a small number of threads that does not vary with the amount
 718 | #   of thrift clients (and thus scales well to many clients). The rpc requests are still
 719 | #   synchronous (one thread per active request). If hsha is selected then it is essential
 720 | #   that rpc_max_threads is changed from the default value of unlimited.
 721 | #
 722 | # The default is sync because on Windows hsha is about 30% slower.  On Linux,
 723 | # sync/hsha performance is about the same, with hsha of course using less memory.
 724 | #
 725 | # Alternatively,  can provide your own RPC server by providing the fully-qualified class name
 726 | # of an o.a.c.t.TServerFactory that can create an instance of it.
 727 | rpc_server_type: sync
 728 | 
 729 | # Uncomment rpc_min|max_thread to set request pool size limits.
 730 | #
 731 | # Regardless of your choice of RPC server (see above), the number of maximum requests in the
 732 | # RPC thread pool dictates how many concurrent requests are possible (but if you are using the sync
 733 | # RPC server, it also dictates the number of clients that can be connected at all).
 734 | #
 735 | # The default is unlimited and thus provides no protection against clients overwhelming the server. You are
 736 | # encouraged to set a maximum that makes sense for you in production, but do keep in mind that
 737 | # rpc_max_threads represents the maximum number of client requests this server may execute concurrently.
 738 | #
 739 | # rpc_min_threads: 16
 740 | # rpc_max_threads: 2048
 741 | 
 742 | # uncomment to set socket buffer sizes on rpc connections
 743 | # rpc_send_buff_size_in_bytes:
 744 | # rpc_recv_buff_size_in_bytes:
 745 | 
 746 | # Uncomment to set socket buffer size for internode communication
 747 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 748 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 749 | # See also:
 750 | # /proc/sys/net/core/wmem_max
 751 | # /proc/sys/net/core/rmem_max
 752 | # /proc/sys/net/ipv4/tcp_wmem
 753 | # /proc/sys/net/ipv4/tcp_wmem
 754 | # and 'man tcp'
 755 | # internode_send_buff_size_in_bytes:
 756 | 
 757 | # Uncomment to set socket buffer size for internode communication
 758 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 759 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 760 | # internode_recv_buff_size_in_bytes:
 761 | 
 762 | # Frame size for thrift (maximum message length).
 763 | thrift_framed_transport_size_in_mb: 15
 764 | 
 765 | # Set to true to have Cassandra create a hard link to each sstable
 766 | # flushed or streamed locally in a backups/ subdirectory of the
 767 | # keyspace data.  Removing these links is the operator's
 768 | # responsibility.
 769 | incremental_backups: false
 770 | 
 771 | # Whether or not to take a snapshot before each compaction.  Be
 772 | # careful using this option, since Cassandra won't clean up the
 773 | # snapshots for you.  Mostly useful if you're paranoid when there
 774 | # is a data format change.
 775 | snapshot_before_compaction: false
 776 | 
 777 | # Whether or not a snapshot is taken of the data before keyspace truncation
 778 | # or dropping of column families. The STRONGLY advised default of true 
 779 | # should be used to provide data safety. If you set this flag to false, you will
 780 | # lose data on truncation or drop.
 781 | auto_snapshot: true
 782 | 
 783 | # Granularity of the collation index of rows within a partition.
 784 | # Increase if your rows are large, or if you have a very large
 785 | # number of rows per partition.  The competing goals are these:
 786 | #
 787 | # - a smaller granularity means more index entries are generated
 788 | #   and looking up rows withing the partition by collation column
 789 | #   is faster
 790 | # - but, Cassandra will keep the collation index in memory for hot
 791 | #   rows (as part of the key cache), so a larger granularity means
 792 | #   you can cache more hot rows
 793 | column_index_size_in_kb: 64
 794 | 
 795 | # Per sstable indexed key cache entries (the collation index in memory
 796 | # mentioned above) exceeding this size will not be held on heap.
 797 | # This means that only partition information is held on heap and the
 798 | # index entries are read from disk.
 799 | #
 800 | # Note that this size refers to the size of the
 801 | # serialized index information and not the size of the partition.
 802 | column_index_cache_size_in_kb: 2
 803 | 
 804 | # Number of simultaneous compactions to allow, NOT including
 805 | # validation "compactions" for anti-entropy repair.  Simultaneous
 806 | # compactions can help preserve read performance in a mixed read/write
 807 | # workload, by mitigating the tendency of small sstables to accumulate
 808 | # during a single long running compactions. The default is usually
 809 | # fine and if you experience problems with compaction running too
 810 | # slowly or too fast, you should look at
 811 | # compaction_throughput_mb_per_sec first.
 812 | #
 813 | # concurrent_compactors defaults to the smaller of (number of disks,
 814 | # number of cores), with a minimum of 2 and a maximum of 8.
 815 | # 
 816 | # If your data directories are backed by SSD, you should increase this
 817 | # to the number of cores.
 818 | #concurrent_compactors: 1
 819 | 
 820 | # Throttles compaction to the given total throughput across the entire
 821 | # system. The faster you insert data, the faster you need to compact in
 822 | # order to keep the sstable count down, but in general, setting this to
 823 | # 16 to 32 times the rate you are inserting data is more than sufficient.
 824 | # Setting this to 0 disables throttling. Note that this account for all types
 825 | # of compaction, including validation compaction.
 826 | compaction_throughput_mb_per_sec: 16
 827 | 
 828 | # When compacting, the replacement sstable(s) can be opened before they
 829 | # are completely written, and used in place of the prior sstables for
 830 | # any range that has been written. This helps to smoothly transfer reads 
 831 | # between the sstables, reducing page cache churn and keeping hot rows hot
 832 | sstable_preemptive_open_interval_in_mb: 50
 833 | 
 834 | # Throttles all outbound streaming file transfers on this node to the
 835 | # given total throughput in Mbps. This is necessary because Cassandra does
 836 | # mostly sequential IO when streaming data during bootstrap or repair, which
 837 | # can lead to saturating the network connection and degrading rpc performance.
 838 | # When unset, the default is 200 Mbps or 25 MB/s.
 839 | # stream_throughput_outbound_megabits_per_sec: 200
 840 | 
 841 | # Throttles all streaming file transfer between the datacenters,
 842 | # this setting allows users to throttle inter dc stream throughput in addition
 843 | # to throttling all network stream traffic as configured with
 844 | # stream_throughput_outbound_megabits_per_sec
 845 | # When unset, the default is 200 Mbps or 25 MB/s
 846 | # inter_dc_stream_throughput_outbound_megabits_per_sec: 200
 847 | 
 848 | # How long the coordinator should wait for read operations to complete
 849 | read_request_timeout_in_ms: 5000
 850 | # How long the coordinator should wait for seq or index scans to complete
 851 | range_request_timeout_in_ms: 10000
 852 | # How long the coordinator should wait for writes to complete
 853 | write_request_timeout_in_ms: 2000
 854 | # How long the coordinator should wait for counter writes to complete
 855 | counter_write_request_timeout_in_ms: 5000
 856 | # How long a coordinator should continue to retry a CAS operation
 857 | # that contends with other proposals for the same row
 858 | cas_contention_timeout_in_ms: 1000
 859 | # How long the coordinator should wait for truncates to complete
 860 | # (This can be much longer, because unless auto_snapshot is disabled
 861 | # we need to flush first so we can snapshot before removing the data.)
 862 | truncate_request_timeout_in_ms: 60000
 863 | # The default timeout for other, miscellaneous operations
 864 | request_timeout_in_ms: 10000
 865 | 
 866 | # How long before a node logs slow queries. Select queries that take longer than
 867 | # this timeout to execute, will generate an aggregated log message, so that slow queries
 868 | # can be identified. Set this value to zero to disable slow query logging.
 869 | slow_query_log_timeout_in_ms: 500
 870 | 
 871 | # Enable operation timeout information exchange between nodes to accurately
 872 | # measure request timeouts.  If disabled, replicas will assume that requests
 873 | # were forwarded to them instantly by the coordinator, which means that
 874 | # under overload conditions we will waste that much extra time processing 
 875 | # already-timed-out requests.
 876 | #
 877 | # Warning: before enabling this property make sure to ntp is installed
 878 | # and the times are synchronized between the nodes.
 879 | cross_node_timeout: false
 880 | 
 881 | # Set keep-alive period for streaming
 882 | # This node will send a keep-alive message periodically with this period.
 883 | # If the node does not receive a keep-alive message from the peer for
 884 | # 2 keep-alive cycles the stream session times out and fail
 885 | # Default value is 300s (5 minutes), which means stalled stream
 886 | # times out in 10 minutes by default
 887 | # streaming_keep_alive_period_in_secs: 300
 888 | 
 889 | # phi value that must be reached for a host to be marked down.
 890 | # most users should never need to adjust this.
 891 | # phi_convict_threshold: 8
 892 | 
 893 | # endpoint_snitch -- Set this to a class that implements
 894 | # IEndpointSnitch.  The snitch has two functions:
 895 | #
 896 | # - it teaches Cassandra enough about your network topology to route
 897 | #   requests efficiently
 898 | # - it allows Cassandra to spread replicas around your cluster to avoid
 899 | #   correlated failures. It does this by grouping machines into
 900 | #   "datacenters" and "racks."  Cassandra will do its best not to have
 901 | #   more than one replica on the same "rack" (which may not actually
 902 | #   be a physical location)
 903 | #
 904 | # CASSANDRA WILL NOT ALLOW YOU TO SWITCH TO AN INCOMPATIBLE SNITCH
 905 | # ONCE DATA IS INSERTED INTO THE CLUSTER.  This would cause data loss.
 906 | # This means that if you start with the default SimpleSnitch, which
 907 | # locates every node on "rack1" in "datacenter1", your only options
 908 | # if you need to add another datacenter are GossipingPropertyFileSnitch
 909 | # (and the older PFS).  From there, if you want to migrate to an
 910 | # incompatible snitch like Ec2Snitch you can do it by adding new nodes
 911 | # under Ec2Snitch (which will locate them in a new "datacenter") and
 912 | # decommissioning the old ones.
 913 | #
 914 | # Out of the box, Cassandra provides:
 915 | #
 916 | # SimpleSnitch:
 917 | #    Treats Strategy order as proximity. This can improve cache
 918 | #    locality when disabling read repair.  Only appropriate for
 919 | #    single-datacenter deployments.
 920 | #
 921 | # GossipingPropertyFileSnitch
 922 | #    This should be your go-to snitch for production use.  The rack
 923 | #    and datacenter for the local node are defined in
 924 | #    cassandra-rackdc.properties and propagated to other nodes via
 925 | #    gossip.  If cassandra-topology.properties exists, it is used as a
 926 | #    fallback, allowing migration from the PropertyFileSnitch.
 927 | #
 928 | # PropertyFileSnitch:
 929 | #    Proximity is determined by rack and data center, which are
 930 | #    explicitly configured in cassandra-topology.properties.
 931 | #
 932 | # Ec2Snitch:
 933 | #    Appropriate for EC2 deployments in a single Region. Loads Region
 934 | #    and Availability Zone information from the EC2 API. The Region is
 935 | #    treated as the datacenter, and the Availability Zone as the rack.
 936 | #    Only private IPs are used, so this will not work across multiple
 937 | #    Regions.
 938 | #
 939 | # Ec2MultiRegionSnitch:
 940 | #    Uses public IPs as broadcast_address to allow cross-region
 941 | #    connectivity.  (Thus, you should set seed addresses to the public
 942 | #    IP as well.) You will need to open the storage_port or
 943 | #    ssl_storage_port on the public IP firewall.  (For intra-Region
 944 | #    traffic, Cassandra will switch to the private IP after
 945 | #    establishing a connection.)
 946 | #
 947 | # RackInferringSnitch:
 948 | #    Proximity is determined by rack and data center, which are
 949 | #    assumed to correspond to the 3rd and 2nd octet of each node's IP
 950 | #    address, respectively.  Unless this happens to match your
 951 | #    deployment conventions, this is best used as an example of
 952 | #    writing a custom Snitch class and is provided in that spirit.
 953 | #
 954 | # You can use a custom Snitch by setting this to the full class name
 955 | # of the snitch, which will be assumed to be on your classpath.
 956 | endpoint_snitch: SimpleSnitch
 957 | 
 958 | # controls how often to perform the more expensive part of host score
 959 | # calculation
 960 | dynamic_snitch_update_interval_in_ms: 100 
 961 | # controls how often to reset all host scores, allowing a bad host to
 962 | # possibly recover
 963 | dynamic_snitch_reset_interval_in_ms: 600000
 964 | # if set greater than zero and read_repair_chance is < 1.0, this will allow
 965 | # 'pinning' of replicas to hosts in order to increase cache capacity.
 966 | # The badness threshold will control how much worse the pinned host has to be
 967 | # before the dynamic snitch will prefer other replicas over it.  This is
 968 | # expressed as a double which represents a percentage.  Thus, a value of
 969 | # 0.2 means Cassandra would continue to prefer the static snitch values
 970 | # until the pinned host was 20% worse than the fastest.
 971 | dynamic_snitch_badness_threshold: 0.1
 972 | 
 973 | # request_scheduler -- Set this to a class that implements
 974 | # RequestScheduler, which will schedule incoming client requests
 975 | # according to the specific policy. This is useful for multi-tenancy
 976 | # with a single Cassandra cluster.
 977 | # NOTE: This is specifically for requests from the client and does
 978 | # not affect inter node communication.
 979 | # org.apache.cassandra.scheduler.NoScheduler - No scheduling takes place
 980 | # org.apache.cassandra.scheduler.RoundRobinScheduler - Round robin of
 981 | # client requests to a node with a separate queue for each
 982 | # request_scheduler_id. The scheduler is further customized by
 983 | # request_scheduler_options as described below.
 984 | request_scheduler: org.apache.cassandra.scheduler.NoScheduler
 985 | 
 986 | # Scheduler Options vary based on the type of scheduler
 987 | #
 988 | # NoScheduler
 989 | #   Has no options
 990 | #
 991 | # RoundRobin
 992 | #   throttle_limit
 993 | #     The throttle_limit is the number of in-flight
 994 | #     requests per client.  Requests beyond 
 995 | #     that limit are queued up until
 996 | #     running requests can complete.
 997 | #     The value of 80 here is twice the number of
 998 | #     concurrent_reads + concurrent_writes.
 999 | #   default_weight
1000 | #     default_weight is optional and allows for
1001 | #     overriding the default which is 1.
1002 | #   weights
1003 | #     Weights are optional and will default to 1 or the
1004 | #     overridden default_weight. The weight translates into how
1005 | #     many requests are handled during each turn of the
1006 | #     RoundRobin, based on the scheduler id.
1007 | #
1008 | # request_scheduler_options:
1009 | #    throttle_limit: 80
1010 | #    default_weight: 5
1011 | #    weights:
1012 | #      Keyspace1: 1
1013 | #      Keyspace2: 5
1014 | 
1015 | # request_scheduler_id -- An identifier based on which to perform
1016 | # the request scheduling. Currently the only valid option is keyspace.
1017 | # request_scheduler_id: keyspace
1018 | 
1019 | # Enable or disable inter-node encryption
1020 | # JVM defaults for supported SSL socket protocols and cipher suites can
1021 | # be replaced using custom encryption options. This is not recommended
1022 | # unless you have policies in place that dictate certain settings, or
1023 | # need to disable vulnerable ciphers or protocols in case the JVM cannot
1024 | # be updated.
1025 | # FIPS compliant settings can be configured at JVM level and should not
1026 | # involve changing encryption settings here:
1027 | # https://docs.oracle.com/javase/8/docs/technotes/guides/security/jsse/FIPS.html
1028 | # *NOTE* No custom encryption options are enabled at the moment
1029 | # The available internode options are : all, none, dc, rack
1030 | #
1031 | # If set to dc cassandra will encrypt the traffic between the DCs
1032 | # If set to rack cassandra will encrypt the traffic between the racks
1033 | #
1034 | # The passwords used in these options must match the passwords used when generating
1035 | # the keystore and truststore.  For instructions on generating these files, see:
1036 | # http://download.oracle.com/javase/6/docs/technotes/guides/security/jsse/JSSERefGuide.html#CreateKeystore
1037 | #
1038 | server_encryption_options:
1039 |     internode_encryption: none
1040 |     keystore: conf/.keystore
1041 |     keystore_password: cassandra
1042 |     truststore: conf/.truststore
1043 |     truststore_password: cassandra
1044 |     # More advanced defaults below:
1045 |     # protocol: TLS
1046 |     # algorithm: SunX509
1047 |     # store_type: JKS
1048 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1049 |     # require_client_auth: false
1050 |     # require_endpoint_verification: false
1051 | 
1052 | # enable or disable client/server encryption.
1053 | client_encryption_options:
1054 |     enabled: false
1055 |     # If enabled and optional is set to true encrypted and unencrypted connections are handled.
1056 |     optional: false
1057 |     keystore: conf/.keystore
1058 |     keystore_password: cassandra
1059 |     # require_client_auth: false
1060 |     # Set trustore and truststore_password if require_client_auth is true
1061 |     # truststore: conf/.truststore
1062 |     # truststore_password: cassandra
1063 |     # More advanced defaults below:
1064 |     # protocol: TLS
1065 |     # algorithm: SunX509
1066 |     # store_type: JKS
1067 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1068 | 
1069 | # internode_compression controls whether traffic between nodes is
1070 | # compressed.
1071 | # Can be:
1072 | #
1073 | # all
1074 | #   all traffic is compressed
1075 | #
1076 | # dc
1077 | #   traffic between different datacenters is compressed
1078 | #
1079 | # none
1080 | #   nothing is compressed.
1081 | internode_compression: dc
1082 | 
1083 | # Enable or disable tcp_nodelay for inter-dc communication.
1084 | # Disabling it will result in larger (but fewer) network packets being sent,
1085 | # reducing overhead from the TCP protocol itself, at the cost of increasing
1086 | # latency if you block for cross-datacenter responses.
1087 | inter_dc_tcp_nodelay: false
1088 | 
1089 | # TTL for different trace types used during logging of the repair process.
1090 | tracetype_query_ttl: 86400
1091 | tracetype_repair_ttl: 604800
1092 | 
1093 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1094 | # This threshold can be adjusted to minimize logging if necessary
1095 | # gc_log_threshold_in_ms: 200
1096 | 
1097 | # If unset, all GC Pauses greater than gc_log_threshold_in_ms will log at
1098 | # INFO level
1099 | # UDFs (user defined functions) are disabled by default.
1100 | # As of Cassandra 3.0 there is a sandbox in place that should prevent execution of evil code.
1101 | enable_user_defined_functions: false
1102 | 
1103 | # Enables scripted UDFs (JavaScript UDFs).
1104 | # Java UDFs are always enabled, if enable_user_defined_functions is true.
1105 | # Enable this option to be able to use UDFs with "language javascript" or any custom JSR-223 provider.
1106 | # This option has no effect, if enable_user_defined_functions is false.
1107 | enable_scripted_user_defined_functions: false
1108 | 
1109 | # The default Windows kernel timer and scheduling resolution is 15.6ms for power conservation.
1110 | # Lowering this value on Windows can provide much tighter latency and better throughput, however
1111 | # some virtualized environments may see a negative performance impact from changing this setting
1112 | # below their system default. The sysinternals 'clockres' tool can confirm your system's default
1113 | # setting.
1114 | windows_timer_interval: 1
1115 | 
1116 | 
1117 | # Enables encrypting data at-rest (on disk). Different key providers can be plugged in, but the default reads from
1118 | # a JCE-style keystore. A single keystore can hold multiple keys, but the one referenced by
1119 | # the "key_alias" is the only key that will be used for encrypt opertaions; previously used keys
1120 | # can still (and should!) be in the keystore and will be used on decrypt operations
1121 | # (to handle the case of key rotation).
1122 | #
1123 | # It is strongly recommended to download and install Java Cryptography Extension (JCE)
1124 | # Unlimited Strength Jurisdiction Policy Files for your version of the JDK.
1125 | # (current link: http://www.oracle.com/technetwork/java/javase/downloads/jce8-download-2133166.html)
1126 | #
1127 | # Currently, only the following file types are supported for transparent data encryption, although
1128 | # more are coming in future cassandra releases: commitlog, hints
1129 | transparent_data_encryption_options:
1130 |     enabled: false
1131 |     chunk_length_kb: 64
1132 |     cipher: AES/CBC/PKCS5Padding
1133 |     key_alias: testing:1
1134 |     # CBC IV length for AES needs to be 16 bytes (which is also the default size)
1135 |     # iv_length: 16
1136 |     key_provider: 
1137 |       - class_name: org.apache.cassandra.security.JKSKeyProvider
1138 |         parameters: 
1139 |           - keystore: conf/.keystore
1140 |             keystore_password: cassandra
1141 |             store_type: JCEKS
1142 |             key_password: cassandra
1143 | 
1144 | 
1145 | #####################
1146 | # SAFETY THRESHOLDS #
1147 | #####################
1148 | 
1149 | # When executing a scan, within or across a partition, we need to keep the
1150 | # tombstones seen in memory so we can return them to the coordinator, which
1151 | # will use them to make sure other replicas also know about the deleted rows.
1152 | # With workloads that generate a lot of tombstones, this can cause performance
1153 | # problems and even exaust the server heap.
1154 | # (http://www.datastax.com/dev/blog/cassandra-anti-patterns-queues-and-queue-like-datasets)
1155 | # Adjust the thresholds here if you understand the dangers and want to
1156 | # scan more tombstones anyway.  These thresholds may also be adjusted at runtime
1157 | # using the StorageService mbean.
1158 | tombstone_warn_threshold: 1000
1159 | tombstone_failure_threshold: 100000
1160 | 
1161 | # Log WARN on any multiple-partition batch size exceeding this value. 5kb per batch by default.
1162 | # Caution should be taken on increasing the size of this threshold as it can lead to node instability.
1163 | batch_size_warn_threshold_in_kb: 5
1164 | 
1165 | # Fail any multiple-partition batch exceeding this value. 50kb (10x warn threshold) by default.
1166 | batch_size_fail_threshold_in_kb: 50
1167 | 
1168 | # Log WARN on any batches not of type LOGGED than span across more partitions than this limit
1169 | unlogged_batch_across_partitions_warn_threshold: 10
1170 | 
1171 | # Log a warning when compacting partitions larger than this value
1172 | compaction_large_partition_warning_threshold_mb: 100
1173 | 
1174 | # GC Pauses greater than gc_warn_threshold_in_ms will be logged at WARN level
1175 | # Adjust the threshold based on your application throughput requirement
1176 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1177 | gc_warn_threshold_in_ms: 1000
1178 | 
1179 | # Maximum size of any value in SSTables. Safety measure to detect SSTable corruption
1180 | # early. Any value size larger than this threshold will result into marking an SSTable
1181 | # as corrupted.
1182 | # max_value_size_in_mb: 256
1183 | 
1184 | # Back-pressure settings #
1185 | # If enabled, the coordinator will apply the back-pressure strategy specified below to each mutation
1186 | # sent to replicas, with the aim of reducing pressure on overloaded replicas.
1187 | back_pressure_enabled: false
1188 | # The back-pressure strategy applied.
1189 | # The default implementation, RateBasedBackPressure, takes three arguments:
1190 | # high ratio, factor, and flow type, and uses the ratio between incoming mutation responses and outgoing mutation requests.
1191 | # If below high ratio, outgoing mutations are rate limited according to the incoming rate decreased by the given factor;
1192 | # if above high ratio, the rate limiting is increased by the given factor;
1193 | # such factor is usually best configured between 1 and 10, use larger values for a faster recovery
1194 | # at the expense of potentially more dropped mutations;
1195 | # the rate limiting is applied according to the flow type: if FAST, it's rate limited at the speed of the fastest replica,
1196 | # if SLOW at the speed of the slowest one.
1197 | # New strategies can be added. Implementors need to implement org.apache.cassandra.net.BackpressureStrategy and
1198 | # provide a public constructor accepting a Map<String, Object>.
1199 | back_pressure_strategy:
1200 |     - class_name: org.apache.cassandra.net.RateBasedBackPressure
1201 |       parameters:
1202 |         - high_ratio: 0.90
1203 |           factor: 5
1204 |           flow: FAST
1205 | 
1206 | # Coalescing Strategies #
1207 | # Coalescing multiples messages turns out to significantly boost message processing throughput (think doubling or more).
1208 | # On bare metal, the floor for packet processing throughput is high enough that many applications won't notice, but in
1209 | # virtualized environments, the point at which an application can be bound by network packet processing can be
1210 | # surprisingly low compared to the throughput of task processing that is possible inside a VM. It's not that bare metal
1211 | # doesn't benefit from coalescing messages, it's that the number of packets a bare metal network interface can process
1212 | # is sufficient for many applications such that no load starvation is experienced even without coalescing.
1213 | # There are other benefits to coalescing network messages that are harder to isolate with a simple metric like messages
1214 | # per second. By coalescing multiple tasks together, a network thread can process multiple messages for the cost of one
1215 | # trip to read from a socket, and all the task submission work can be done at the same time reducing context switching
1216 | # and increasing cache friendliness of network message processing.
1217 | # See CASSANDRA-8692 for details.
1218 | 
1219 | # Strategy to use for coalescing messages in OutboundTcpConnection.
1220 | # Can be fixed, movingaverage, timehorizon, disabled (default).
1221 | # You can also specify a subclass of CoalescingStrategies.CoalescingStrategy by name.
1222 | # otc_coalescing_strategy: DISABLED
1223 | 
1224 | # How many microseconds to wait for coalescing. For fixed strategy this is the amount of time after the first
1225 | # message is received before it will be sent with any accompanying messages. For moving average this is the
1226 | # maximum amount of time that will be waited as well as the interval at which messages must arrive on average
1227 | # for coalescing to be enabled.
1228 | # otc_coalescing_window_us: 200
1229 | 
1230 | # Do not try to coalesce messages if we already got that many messages. This should be more than 2 and less than 128.
1231 | # otc_coalescing_enough_coalesced_messages: 8
1232 | 
1233 | # How many milliseconds to wait between two expiration runs on the backlog (queue) of the OutboundTcpConnection.
1234 | # Expiration is done if messages are piling up in the backlog. Droppable messages are expired to free the memory
1235 | # taken by expired messages. The interval should be between 0 and 1000, and in most installations the default value
1236 | # will be appropriate. A smaller value could potentially expire messages slightly sooner at the expense of more CPU
1237 | # time and queue contention while iterating the backlog of messages.
1238 | # An interval of 0 disables any wait time, which is the behavior of former Cassandra versions.
1239 | #
1240 | # otc_backlog_expiration_interval_ms: 200
1241 | 


--------------------------------------------------------------------------------
/cassandra-cdc/config/cassandra-2-cdc-tmp.yaml:
--------------------------------------------------------------------------------
   1 | # Cassandra storage config YAML
   2 | 
   3 | # NOTE:
   4 | #   See http://wiki.apache.org/cassandra/StorageConfiguration for
   5 | #   full explanations of configuration directives
   6 | # /NOTE
   7 | 
   8 | # The name of the cluster. This is mainly used to prevent machines in
   9 | # one logical cluster from joining another.
  10 | cluster_name: 'Test Cluster'
  11 | 
  12 | # This defines the number of tokens randomly assigned to this node on the ring
  13 | # The more tokens, relative to other nodes, the larger the proportion of data
  14 | # that this node will store. You probably want all nodes to have the same number
  15 | # of tokens assuming they have equal hardware capability.
  16 | #
  17 | # If you leave this unspecified, Cassandra will use the default of 1 token for legacy compatibility,
  18 | # and will use the initial_token as described below.
  19 | #
  20 | # Specifying initial_token will override this setting on the node's initial start,
  21 | # on subsequent starts, this setting will apply even if initial token is set.
  22 | #
  23 | # If you already have a cluster with 1 token per node, and wish to migrate to 
  24 | # multiple tokens per node, see http://wiki.apache.org/cassandra/Operations
  25 | num_tokens: 256
  26 | 
  27 | # Triggers automatic allocation of num_tokens tokens for this node. The allocation
  28 | # algorithm attempts to choose tokens in a way that optimizes replicated load over
  29 | # the nodes in the datacenter for the replication strategy used by the specified
  30 | # keyspace.
  31 | #
  32 | # The load assigned to each node will be close to proportional to its number of
  33 | # vnodes.
  34 | #
  35 | # Only supported with the Murmur3Partitioner.
  36 | # allocate_tokens_for_keyspace: KEYSPACE
  37 | 
  38 | # initial_token allows you to specify tokens manually.  While you can use it with
  39 | # vnodes (num_tokens > 1, above) -- in which case you should provide a 
  40 | # comma-separated list -- it's primarily used when adding nodes to legacy clusters 
  41 | # that do not have vnodes enabled.
  42 | # initial_token:
  43 | 
  44 | # See http://wiki.apache.org/cassandra/HintedHandoff
  45 | # May either be "true" or "false" to enable globally
  46 | hinted_handoff_enabled: true
  47 | 
  48 | # When hinted_handoff_enabled is true, a black list of data centers that will not
  49 | # perform hinted handoff
  50 | # hinted_handoff_disabled_datacenters:
  51 | #    - DC1
  52 | #    - DC2
  53 | 
  54 | # this defines the maximum amount of time a dead host will have hints
  55 | # generated.  After it has been dead this long, new hints for it will not be
  56 | # created until it has been seen alive and gone down again.
  57 | max_hint_window_in_ms: 10800000 # 3 hours
  58 | 
  59 | # Maximum throttle in KBs per second, per delivery thread.  This will be
  60 | # reduced proportionally to the number of nodes in the cluster.  (If there
  61 | # are two nodes in the cluster, each delivery thread will use the maximum
  62 | # rate; if there are three, each will throttle to half of the maximum,
  63 | # since we expect two nodes to be delivering hints simultaneously.)
  64 | hinted_handoff_throttle_in_kb: 1024
  65 | 
  66 | # Number of threads with which to deliver hints;
  67 | # Consider increasing this number when you have multi-dc deployments, since
  68 | # cross-dc handoff tends to be slower
  69 | max_hints_delivery_threads: 2
  70 | 
  71 | # Directory where Cassandra should store hints.
  72 | # If not set, the default directory is $CASSANDRA_HOME/data/hints.
  73 | # hints_directory: /var/lib/cassandra/hints
  74 | hints_directory: /tmp/cdc/cassandra-2/hints
  75 | 
  76 | # How often hints should be flushed from the internal buffers to disk.
  77 | # Will *not* trigger fsync.
  78 | hints_flush_period_in_ms: 10000
  79 | 
  80 | # Maximum size for a single hints file, in megabytes.
  81 | max_hints_file_size_in_mb: 128
  82 | 
  83 | # Compression to apply to the hint files. If omitted, hints files
  84 | # will be written uncompressed. LZ4, Snappy, and Deflate compressors
  85 | # are supported.
  86 | #hints_compression:
  87 | #   - class_name: LZ4Compressor
  88 | #     parameters:
  89 | #         -
  90 | 
  91 | # Maximum throttle in KBs per second, total. This will be
  92 | # reduced proportionally to the number of nodes in the cluster.
  93 | batchlog_replay_throttle_in_kb: 1024
  94 | 
  95 | # Authentication backend, implementing IAuthenticator; used to identify users
  96 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthenticator,
  97 | # PasswordAuthenticator}.
  98 | #
  99 | # - AllowAllAuthenticator performs no checks - set it to disable authentication.
 100 | # - PasswordAuthenticator relies on username/password pairs to authenticate
 101 | #   users. It keeps usernames and hashed passwords in system_auth.roles table.
 102 | #   Please increase system_auth keyspace replication factor if you use this authenticator.
 103 | #   If using PasswordAuthenticator, CassandraRoleManager must also be used (see below)
 104 | authenticator: AllowAllAuthenticator
 105 | 
 106 | # Authorization backend, implementing IAuthorizer; used to limit access/provide permissions
 107 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthorizer,
 108 | # CassandraAuthorizer}.
 109 | #
 110 | # - AllowAllAuthorizer allows any action to any user - set it to disable authorization.
 111 | # - CassandraAuthorizer stores permissions in system_auth.role_permissions table. Please
 112 | #   increase system_auth keyspace replication factor if you use this authorizer.
 113 | authorizer: AllowAllAuthorizer
 114 | 
 115 | # Part of the Authentication & Authorization backend, implementing IRoleManager; used
 116 | # to maintain grants and memberships between roles.
 117 | # Out of the box, Cassandra provides org.apache.cassandra.auth.CassandraRoleManager,
 118 | # which stores role information in the system_auth keyspace. Most functions of the
 119 | # IRoleManager require an authenticated login, so unless the configured IAuthenticator
 120 | # actually implements authentication, most of this functionality will be unavailable.
 121 | #
 122 | # - CassandraRoleManager stores role data in the system_auth keyspace. Please
 123 | #   increase system_auth keyspace replication factor if you use this role manager.
 124 | role_manager: CassandraRoleManager
 125 | 
 126 | # Validity period for roles cache (fetching granted roles can be an expensive
 127 | # operation depending on the role manager, CassandraRoleManager is one example)
 128 | # Granted roles are cached for authenticated sessions in AuthenticatedUser and
 129 | # after the period specified here, become eligible for (async) reload.
 130 | # Defaults to 2000, set to 0 to disable caching entirely.
 131 | # Will be disabled automatically for AllowAllAuthenticator.
 132 | roles_validity_in_ms: 2000
 133 | 
 134 | # Refresh interval for roles cache (if enabled).
 135 | # After this interval, cache entries become eligible for refresh. Upon next
 136 | # access, an async reload is scheduled and the old value returned until it
 137 | # completes. If roles_validity_in_ms is non-zero, then this must be
 138 | # also.
 139 | # Defaults to the same value as roles_validity_in_ms.
 140 | # roles_update_interval_in_ms: 2000
 141 | 
 142 | # Validity period for permissions cache (fetching permissions can be an
 143 | # expensive operation depending on the authorizer, CassandraAuthorizer is
 144 | # one example). Defaults to 2000, set to 0 to disable.
 145 | # Will be disabled automatically for AllowAllAuthorizer.
 146 | permissions_validity_in_ms: 2000
 147 | 
 148 | # Refresh interval for permissions cache (if enabled).
 149 | # After this interval, cache entries become eligible for refresh. Upon next
 150 | # access, an async reload is scheduled and the old value returned until it
 151 | # completes. If permissions_validity_in_ms is non-zero, then this must be
 152 | # also.
 153 | # Defaults to the same value as permissions_validity_in_ms.
 154 | # permissions_update_interval_in_ms: 2000
 155 | 
 156 | # Validity period for credentials cache. This cache is tightly coupled to
 157 | # the provided PasswordAuthenticator implementation of IAuthenticator. If
 158 | # another IAuthenticator implementation is configured, this cache will not
 159 | # be automatically used and so the following settings will have no effect.
 160 | # Please note, credentials are cached in their encrypted form, so while
 161 | # activating this cache may reduce the number of queries made to the
 162 | # underlying table, it may not  bring a significant reduction in the
 163 | # latency of individual authentication attempts.
 164 | # Defaults to 2000, set to 0 to disable credentials caching.
 165 | credentials_validity_in_ms: 2000
 166 | 
 167 | # Refresh interval for credentials cache (if enabled).
 168 | # After this interval, cache entries become eligible for refresh. Upon next
 169 | # access, an async reload is scheduled and the old value returned until it
 170 | # completes. If credentials_validity_in_ms is non-zero, then this must be
 171 | # also.
 172 | # Defaults to the same value as credentials_validity_in_ms.
 173 | # credentials_update_interval_in_ms: 2000
 174 | 
 175 | # The partitioner is responsible for distributing groups of rows (by
 176 | # partition key) across nodes in the cluster.  You should leave this
 177 | # alone for new clusters.  The partitioner can NOT be changed without
 178 | # reloading all data, so when upgrading you should set this to the
 179 | # same partitioner you were already using.
 180 | #
 181 | # Besides Murmur3Partitioner, partitioners included for backwards
 182 | # compatibility include RandomPartitioner, ByteOrderedPartitioner, and
 183 | # OrderPreservingPartitioner.
 184 | #
 185 | partitioner: org.apache.cassandra.dht.Murmur3Partitioner
 186 | 
 187 | # Directories where Cassandra should store data on disk.  Cassandra
 188 | # will spread data evenly across them, subject to the granularity of
 189 | # the configured compaction strategy.
 190 | # If not set, the default directory is $CASSANDRA_HOME/data/data.
 191 | # data_file_directories:
 192 | #     - /var/lib/cassandra/data
 193 | data_file_directories:
 194 |   - /tmp/cdc/cassandra-2/data
 195 | 
 196 | # commit log.  when running on magnetic HDD, this should be a
 197 | # separate spindle than the data directories.
 198 | # If not set, the default directory is $CASSANDRA_HOME/data/commitlog.
 199 | # commitlog_directory: /var/lib/cassandra/commitlog
 200 | commitlog_directory: /tmp/cdc/cassandra-2/commitlog
 201 | 
 202 | # Enable / disable CDC functionality on a per-node basis. This modifies the logic used
 203 | # for write path allocation rejection (standard: never reject. cdc: reject Mutation
 204 | # containing a CDC-enabled table if at space limit in cdc_raw_directory).
 205 | cdc_enabled: true
 206 | 
 207 | # CommitLogSegments are moved to this directory on flush if cdc_enabled: true and the
 208 | # segment contains mutations for a CDC-enabled table. This should be placed on a
 209 | # separate spindle than the data directories. If not set, the default directory is
 210 | # $CASSANDRA_HOME/data/cdc_raw.
 211 | # cdc_raw_directory: /var/lib/cassandra/cdc_raw
 212 | cdc_raw_directory: /tmp/cdc/cassandra-2/cdc_raw
 213 | 
 214 | # Policy for data disk failures:
 215 | #
 216 | # die
 217 | #   shut down gossip and client transports and kill the JVM for any fs errors or
 218 | #   single-sstable errors, so the node can be replaced.
 219 | #
 220 | # stop_paranoid
 221 | #   shut down gossip and client transports even for single-sstable errors,
 222 | #   kill the JVM for errors during startup.
 223 | #
 224 | # stop
 225 | #   shut down gossip and client transports, leaving the node effectively dead, but
 226 | #   can still be inspected via JMX, kill the JVM for errors during startup.
 227 | #
 228 | # best_effort
 229 | #    stop using the failed disk and respond to requests based on
 230 | #    remaining available sstables.  This means you WILL see obsolete
 231 | #    data at CL.ONE!
 232 | #
 233 | # ignore
 234 | #    ignore fatal errors and let requests fail, as in pre-1.2 Cassandra
 235 | disk_failure_policy: stop
 236 | 
 237 | # Policy for commit disk failures:
 238 | #
 239 | # die
 240 | #   shut down gossip and Thrift and kill the JVM, so the node can be replaced.
 241 | #
 242 | # stop
 243 | #   shut down gossip and Thrift, leaving the node effectively dead, but
 244 | #   can still be inspected via JMX.
 245 | #
 246 | # stop_commit
 247 | #   shutdown the commit log, letting writes collect but
 248 | #   continuing to service reads, as in pre-2.0.5 Cassandra
 249 | #
 250 | # ignore
 251 | #   ignore fatal errors and let the batches fail
 252 | commit_failure_policy: stop
 253 | 
 254 | # Maximum size of the native protocol prepared statement cache
 255 | #
 256 | # Valid values are either "auto" (omitting the value) or a value greater 0.
 257 | #
 258 | # Note that specifying a too large value will result in long running GCs and possbily
 259 | # out-of-memory errors. Keep the value at a small fraction of the heap.
 260 | #
 261 | # If you constantly see "prepared statements discarded in the last minute because
 262 | # cache limit reached" messages, the first step is to investigate the root cause
 263 | # of these messages and check whether prepared statements are used correctly -
 264 | # i.e. use bind markers for variable parts.
 265 | #
 266 | # Do only change the default value, if you really have more prepared statements than
 267 | # fit in the cache. In most cases it is not neccessary to change this value.
 268 | # Constantly re-preparing statements is a performance penalty.
 269 | #
 270 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 271 | prepared_statements_cache_size_mb:
 272 | 
 273 | # Maximum size of the Thrift prepared statement cache
 274 | #
 275 | # If you do not use Thrift at all, it is safe to leave this value at "auto".
 276 | #
 277 | # See description of 'prepared_statements_cache_size_mb' above for more information.
 278 | #
 279 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 280 | thrift_prepared_statements_cache_size_mb:
 281 | 
 282 | # Maximum size of the key cache in memory.
 283 | #
 284 | # Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at the
 285 | # minimum, sometimes more. The key cache is fairly tiny for the amount of
 286 | # time it saves, so it's worthwhile to use it at large numbers.
 287 | # The row cache saves even more time, but must contain the entire row,
 288 | # so it is extremely space-intensive. It's best to only use the
 289 | # row cache if you have hot rows or static rows.
 290 | #
 291 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 292 | #
 293 | # Default value is empty to make it "auto" (min(5% of Heap (in MB), 100MB)). Set to 0 to disable key cache.
 294 | key_cache_size_in_mb:
 295 | 
 296 | # Duration in seconds after which Cassandra should
 297 | # save the key cache. Caches are saved to saved_caches_directory as
 298 | # specified in this configuration file.
 299 | #
 300 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 301 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 302 | # has limited use.
 303 | #
 304 | # Default is 14400 or 4 hours.
 305 | key_cache_save_period: 14400
 306 | 
 307 | # Number of keys from the key cache to save
 308 | # Disabled by default, meaning all keys are going to be saved
 309 | # key_cache_keys_to_save: 100
 310 | 
 311 | # Row cache implementation class name. Available implementations:
 312 | #
 313 | # org.apache.cassandra.cache.OHCProvider
 314 | #   Fully off-heap row cache implementation (default).
 315 | #
 316 | # org.apache.cassandra.cache.SerializingCacheProvider
 317 | #   This is the row cache implementation availabile
 318 | #   in previous releases of Cassandra.
 319 | # row_cache_class_name: org.apache.cassandra.cache.OHCProvider
 320 | 
 321 | # Maximum size of the row cache in memory.
 322 | # Please note that OHC cache implementation requires some additional off-heap memory to manage
 323 | # the map structures and some in-flight memory during operations before/after cache entries can be
 324 | # accounted against the cache capacity. This overhead is usually small compared to the whole capacity.
 325 | # Do not specify more memory that the system can afford in the worst usual situation and leave some
 326 | # headroom for OS block level cache. Do never allow your system to swap.
 327 | #
 328 | # Default value is 0, to disable row caching.
 329 | row_cache_size_in_mb: 0
 330 | 
 331 | # Duration in seconds after which Cassandra should save the row cache.
 332 | # Caches are saved to saved_caches_directory as specified in this configuration file.
 333 | #
 334 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 335 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 336 | # has limited use.
 337 | #
 338 | # Default is 0 to disable saving the row cache.
 339 | row_cache_save_period: 0
 340 | 
 341 | # Number of keys from the row cache to save.
 342 | # Specify 0 (which is the default), meaning all keys are going to be saved
 343 | # row_cache_keys_to_save: 100
 344 | 
 345 | # Maximum size of the counter cache in memory.
 346 | #
 347 | # Counter cache helps to reduce counter locks' contention for hot counter cells.
 348 | # In case of RF = 1 a counter cache hit will cause Cassandra to skip the read before
 349 | # write entirely. With RF > 1 a counter cache hit will still help to reduce the duration
 350 | # of the lock hold, helping with hot counter cell updates, but will not allow skipping
 351 | # the read entirely. Only the local (clock, count) tuple of a counter cell is kept
 352 | # in memory, not the whole counter, so it's relatively cheap.
 353 | #
 354 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 355 | #
 356 | # Default value is empty to make it "auto" (min(2.5% of Heap (in MB), 50MB)). Set to 0 to disable counter cache.
 357 | # NOTE: if you perform counter deletes and rely on low gcgs, you should disable the counter cache.
 358 | counter_cache_size_in_mb:
 359 | 
 360 | # Duration in seconds after which Cassandra should
 361 | # save the counter cache (keys only). Caches are saved to saved_caches_directory as
 362 | # specified in this configuration file.
 363 | #
 364 | # Default is 7200 or 2 hours.
 365 | counter_cache_save_period: 7200
 366 | 
 367 | # Number of keys from the counter cache to save
 368 | # Disabled by default, meaning all keys are going to be saved
 369 | # counter_cache_keys_to_save: 100
 370 | 
 371 | # saved caches
 372 | # If not set, the default directory is $CASSANDRA_HOME/data/saved_caches.
 373 | # saved_caches_directory: /var/lib/cassandra/saved_caches
 374 | 
 375 | # commitlog_sync may be either "periodic" or "batch." 
 376 | # 
 377 | # When in batch mode, Cassandra won't ack writes until the commit log
 378 | # has been fsynced to disk.  It will wait
 379 | # commitlog_sync_batch_window_in_ms milliseconds between fsyncs.
 380 | # This window should be kept short because the writer threads will
 381 | # be unable to do extra work while waiting.  (You may need to increase
 382 | # concurrent_writes for the same reason.)
 383 | #
 384 | # commitlog_sync: batch
 385 | # commitlog_sync_batch_window_in_ms: 2
 386 | #
 387 | # the other option is "periodic" where writes may be acked immediately
 388 | # and the CommitLog is simply synced every commitlog_sync_period_in_ms
 389 | # milliseconds. 
 390 | commitlog_sync: periodic
 391 | commitlog_sync_period_in_ms: 10000
 392 | 
 393 | # The size of the individual commitlog file segments.  A commitlog
 394 | # segment may be archived, deleted, or recycled once all the data
 395 | # in it (potentially from each columnfamily in the system) has been
 396 | # flushed to sstables.
 397 | #
 398 | # The default size is 32, which is almost always fine, but if you are
 399 | # archiving commitlog segments (see commitlog_archiving.properties),
 400 | # then you probably want a finer granularity of archiving; 8 or 16 MB
 401 | # is reasonable.
 402 | # Max mutation size is also configurable via max_mutation_size_in_kb setting in
 403 | # cassandra.yaml. The default is half the size commitlog_segment_size_in_mb * 1024.
 404 | #
 405 | # NOTE: If max_mutation_size_in_kb is set explicitly then commitlog_segment_size_in_mb must
 406 | # be set to at least twice the size of max_mutation_size_in_kb / 1024
 407 | #
 408 | commitlog_segment_size_in_mb: 1
 409 | 
 410 | # Compression to apply to the commit log. If omitted, the commit log
 411 | # will be written uncompressed.  LZ4, Snappy, and Deflate compressors
 412 | # are supported.
 413 | # commitlog_compression:
 414 | #   - class_name: LZ4Compressor
 415 | #     parameters:
 416 | #         -
 417 | 
 418 | # any class that implements the SeedProvider interface and has a
 419 | # constructor that takes a Map<String, String> of parameters will do.
 420 | seed_provider:
 421 |     # Addresses of hosts that are deemed contact points. 
 422 |     # Cassandra nodes use this list of hosts to find each other and learn
 423 |     # the topology of the ring.  You must change this if you are running
 424 |     # multiple nodes!
 425 |     - class_name: org.apache.cassandra.locator.SimpleSeedProvider
 426 |       parameters:
 427 |           # seeds is actually a comma-delimited list of addresses.
 428 |           # Ex: "<ip1>,<ip2>,<ip3>"
 429 |           - seeds: "127.0.0.1"
 430 | 
 431 | # For workloads with more data than can fit in memory, Cassandra's
 432 | # bottleneck will be reads that need to fetch data from
 433 | # disk. "concurrent_reads" should be set to (16 * number_of_drives) in
 434 | # order to allow the operations to enqueue low enough in the stack
 435 | # that the OS and drives can reorder them. Same applies to
 436 | # "concurrent_counter_writes", since counter writes read the current
 437 | # values before incrementing and writing them back.
 438 | #
 439 | # On the other hand, since writes are almost never IO bound, the ideal
 440 | # number of "concurrent_writes" is dependent on the number of cores in
 441 | # your system; (8 * number_of_cores) is a good rule of thumb.
 442 | concurrent_reads: 32
 443 | concurrent_writes: 32
 444 | concurrent_counter_writes: 32
 445 | 
 446 | # For materialized view writes, as there is a read involved, so this should
 447 | # be limited by the less of concurrent reads or concurrent writes.
 448 | concurrent_materialized_view_writes: 32
 449 | 
 450 | # Maximum memory to use for sstable chunk cache and buffer pooling.
 451 | # 32MB of this are reserved for pooling buffers, the rest is used as an
 452 | # cache that holds uncompressed sstable chunks.
 453 | # Defaults to the smaller of 1/4 of heap or 512MB. This pool is allocated off-heap,
 454 | # so is in addition to the memory allocated for heap. The cache also has on-heap
 455 | # overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size
 456 | # if the default 64k chunk size is used).
 457 | # Memory is only allocated when needed.
 458 | # file_cache_size_in_mb: 512
 459 | 
 460 | # Flag indicating whether to allocate on or off heap when the sstable buffer
 461 | # pool is exhausted, that is when it has exceeded the maximum memory
 462 | # file_cache_size_in_mb, beyond which it will not cache buffers but allocate on request.
 463 | 
 464 | # buffer_pool_use_heap_if_exhausted: true
 465 | 
 466 | # The strategy for optimizing disk read
 467 | # Possible values are:
 468 | # ssd (for solid state disks, the default)
 469 | # spinning (for spinning disks)
 470 | # disk_optimization_strategy: ssd
 471 | 
 472 | # Total permitted memory to use for memtables. Cassandra will stop
 473 | # accepting writes when the limit is exceeded until a flush completes,
 474 | # and will trigger a flush based on memtable_cleanup_threshold
 475 | # If omitted, Cassandra will set both to 1/4 the size of the heap.
 476 | # memtable_heap_space_in_mb: 2048
 477 | # memtable_offheap_space_in_mb: 2048
 478 | 
 479 | # memtable_cleanup_threshold is deprecated. The default calculation
 480 | # is the only reasonable choice. See the comments on  memtable_flush_writers
 481 | # for more information.
 482 | #
 483 | # Ratio of occupied non-flushing memtable size to total permitted size
 484 | # that will trigger a flush of the largest memtable. Larger mct will
 485 | # mean larger flushes and hence less compaction, but also less concurrent
 486 | # flush activity which can make it difficult to keep your disks fed
 487 | # under heavy write load.
 488 | #
 489 | # memtable_cleanup_threshold defaults to 1 / (memtable_flush_writers + 1)
 490 | # memtable_cleanup_threshold: 0.11
 491 | 
 492 | # Specify the way Cassandra allocates and manages memtable memory.
 493 | # Options are:
 494 | #
 495 | # heap_buffers
 496 | #   on heap nio buffers
 497 | #
 498 | # offheap_buffers
 499 | #   off heap (direct) nio buffers
 500 | #
 501 | # offheap_objects
 502 | #    off heap objects
 503 | memtable_allocation_type: heap_buffers
 504 | 
 505 | # Total space to use for commit logs on disk.
 506 | #
 507 | # If space gets above this value, Cassandra will flush every dirty CF
 508 | # in the oldest segment and remove it.  So a small total commitlog space
 509 | # will tend to cause more flush activity on less-active columnfamilies.
 510 | #
 511 | # The default value is the smaller of 8192, and 1/4 of the total space
 512 | # of the commitlog volume.
 513 | #
 514 | # commitlog_total_space_in_mb: 8192
 515 | commitlog_total_space_in_mb: 16
 516 | 
 517 | # This sets the number of memtable flush writer threads per disk
 518 | # as well as the total number of memtables that can be flushed concurrently.
 519 | # These are generally a combination of compute and IO bound.
 520 | #
 521 | # Memtable flushing is more CPU efficient than memtable ingest and a single thread
 522 | # can keep up with the ingest rate of a whole server on a single fast disk
 523 | # until it temporarily becomes IO bound under contention typically with compaction.
 524 | # At that point you need multiple flush threads. At some point in the future
 525 | # it may become CPU bound all the time.
 526 | #
 527 | # You can tell if flushing is falling behind using the MemtablePool.BlockedOnAllocation
 528 | # metric which should be 0, but will be non-zero if threads are blocked waiting on flushing
 529 | # to free memory.
 530 | #
 531 | # memtable_flush_writers defaults to two for a single data directory.
 532 | # This means that two  memtables can be flushed concurrently to the single data directory.
 533 | # If you have multiple data directories the default is one memtable flushing at a time
 534 | # but the flush will use a thread per data directory so you will get two or more writers.
 535 | #
 536 | # Two is generally enough to flush on a fast disk [array] mounted as a single data directory.
 537 | # Adding more flush writers will result in smaller more frequent flushes that introduce more
 538 | # compaction overhead.
 539 | #
 540 | # There is a direct tradeoff between number of memtables that can be flushed concurrently
 541 | # and flush size and frequency. More is not better you just need enough flush writers
 542 | # to never stall waiting for flushing to free memory.
 543 | #
 544 | #memtable_flush_writers: 2
 545 | 
 546 | # Total space to use for change-data-capture logs on disk.
 547 | #
 548 | # If space gets above this value, Cassandra will throw WriteTimeoutException
 549 | # on Mutations including tables with CDC enabled. A CDCCompactor is responsible
 550 | # for parsing the raw CDC logs and deleting them when parsing is completed.
 551 | #
 552 | # The default value is the min of 4096 mb and 1/8th of the total space
 553 | # of the drive where cdc_raw_directory resides.
 554 | # cdc_total_space_in_mb: 4096
 555 | cdc_total_space_in_mb: 4096
 556 | 
 557 | # When we hit our cdc_raw limit and the CDCCompactor is either running behind
 558 | # or experiencing backpressure, we check at the following interval to see if any
 559 | # new space for cdc-tracked tables has been made available. Default to 250ms
 560 | # cdc_free_space_check_interval_ms: 250
 561 | cdc_free_space_check_interval_ms: 250
 562 | 
 563 | # A fixed memory pool size in MB for for SSTable index summaries. If left
 564 | # empty, this will default to 5% of the heap size. If the memory usage of
 565 | # all index summaries exceeds this limit, SSTables with low read rates will
 566 | # shrink their index summaries in order to meet this limit.  However, this
 567 | # is a best-effort process. In extreme conditions Cassandra may need to use
 568 | # more than this amount of memory.
 569 | index_summary_capacity_in_mb:
 570 | 
 571 | # How frequently index summaries should be resampled.  This is done
 572 | # periodically to redistribute memory from the fixed-size pool to sstables
 573 | # proportional their recent read rates.  Setting to -1 will disable this
 574 | # process, leaving existing index summaries at their current sampling level.
 575 | index_summary_resize_interval_in_minutes: 60
 576 | 
 577 | # Whether to, when doing sequential writing, fsync() at intervals in
 578 | # order to force the operating system to flush the dirty
 579 | # buffers. Enable this to avoid sudden dirty buffer flushing from
 580 | # impacting read latencies. Almost always a good idea on SSDs; not
 581 | # necessarily on platters.
 582 | trickle_fsync: false
 583 | trickle_fsync_interval_in_kb: 10240
 584 | 
 585 | # TCP port, for commands and data
 586 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 587 | storage_port: 7000
 588 | 
 589 | # SSL port, for encrypted communication.  Unused unless enabled in
 590 | # encryption_options
 591 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 592 | ssl_storage_port: 7001
 593 | 
 594 | # Address or interface to bind to and tell other Cassandra nodes to connect to.
 595 | # You _must_ change this if you want multiple nodes to be able to communicate!
 596 | #
 597 | # Set listen_address OR listen_interface, not both.
 598 | #
 599 | # Leaving it blank leaves it up to InetAddress.getLocalHost(). This
 600 | # will always do the Right Thing _if_ the node is properly configured
 601 | # (hostname, name resolution, etc), and the Right Thing is to use the
 602 | # address associated with the hostname (it might not be).
 603 | #
 604 | # Setting listen_address to 0.0.0.0 is always wrong.
 605 | #
 606 | listen_address: localhost
 607 | 
 608 | # Set listen_address OR listen_interface, not both. Interfaces must correspond
 609 | # to a single address, IP aliasing is not supported.
 610 | # listen_interface: eth0
 611 | 
 612 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 613 | # you can specify which should be chosen using listen_interface_prefer_ipv6. If false the first ipv4
 614 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 615 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 616 | # listen_interface_prefer_ipv6: false
 617 | 
 618 | # Address to broadcast to other Cassandra nodes
 619 | # Leaving this blank will set it to the same value as listen_address
 620 | # broadcast_address: 1.2.3.4
 621 | 
 622 | # When using multiple physical network interfaces, set this
 623 | # to true to listen on broadcast_address in addition to
 624 | # the listen_address, allowing nodes to communicate in both
 625 | # interfaces.
 626 | # Ignore this property if the network configuration automatically
 627 | # routes  between the public and private networks such as EC2.
 628 | # listen_on_broadcast_address: false
 629 | 
 630 | # Internode authentication backend, implementing IInternodeAuthenticator;
 631 | # used to allow/disallow connections from peer nodes.
 632 | # internode_authenticator: org.apache.cassandra.auth.AllowAllInternodeAuthenticator
 633 | 
 634 | # Whether to start the native transport server.
 635 | # Please note that the address on which the native transport is bound is the
 636 | # same as the rpc_address. The port however is different and specified below.
 637 | start_native_transport: true
 638 | # port for the CQL native transport to listen for clients on
 639 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 640 | native_transport_port: 9042
 641 | # Enabling native transport encryption in client_encryption_options allows you to either use
 642 | # encryption for the standard port or to use a dedicated, additional port along with the unencrypted
 643 | # standard native_transport_port.
 644 | # Enabling client encryption and keeping native_transport_port_ssl disabled will use encryption
 645 | # for native_transport_port. Setting native_transport_port_ssl to a different value
 646 | # from native_transport_port will use encryption for native_transport_port_ssl while
 647 | # keeping native_transport_port unencrypted.
 648 | # native_transport_port_ssl: 9142
 649 | # The maximum threads for handling requests when the native transport is used.
 650 | # This is similar to rpc_max_threads though the default differs slightly (and
 651 | # there is no native_transport_min_threads, idle threads will always be stopped
 652 | # after 30 seconds).
 653 | # native_transport_max_threads: 128
 654 | #
 655 | # The maximum size of allowed frame. Frame (requests) larger than this will
 656 | # be rejected as invalid. The default is 256MB. If you're changing this parameter,
 657 | # you may want to adjust max_value_size_in_mb accordingly.
 658 | # native_transport_max_frame_size_in_mb: 256
 659 | 
 660 | # The maximum number of concurrent client connections.
 661 | # The default is -1, which means unlimited.
 662 | # native_transport_max_concurrent_connections: -1
 663 | 
 664 | # The maximum number of concurrent client connections per source ip.
 665 | # The default is -1, which means unlimited.
 666 | # native_transport_max_concurrent_connections_per_ip: -1
 667 | 
 668 | # Whether to start the thrift rpc server.
 669 | start_rpc: false
 670 | 
 671 | # The address or interface to bind the Thrift RPC service and native transport
 672 | # server to.
 673 | #
 674 | # Set rpc_address OR rpc_interface, not both.
 675 | #
 676 | # Leaving rpc_address blank has the same effect as on listen_address
 677 | # (i.e. it will be based on the configured hostname of the node).
 678 | #
 679 | # Note that unlike listen_address, you can specify 0.0.0.0, but you must also
 680 | # set broadcast_rpc_address to a value other than 0.0.0.0.
 681 | #
 682 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 683 | rpc_address: localhost
 684 | 
 685 | # Set rpc_address OR rpc_interface, not both. Interfaces must correspond
 686 | # to a single address, IP aliasing is not supported.
 687 | # rpc_interface: eth1
 688 | 
 689 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 690 | # you can specify which should be chosen using rpc_interface_prefer_ipv6. If false the first ipv4
 691 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 692 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 693 | # rpc_interface_prefer_ipv6: false
 694 | 
 695 | # port for Thrift to listen for clients on
 696 | rpc_port: 9160
 697 | 
 698 | # RPC address to broadcast to drivers and other Cassandra nodes. This cannot
 699 | # be set to 0.0.0.0. If left blank, this will be set to the value of
 700 | # rpc_address. If rpc_address is set to 0.0.0.0, broadcast_rpc_address must
 701 | # be set.
 702 | # broadcast_rpc_address: 1.2.3.4
 703 | 
 704 | # enable or disable keepalive on rpc/native connections
 705 | rpc_keepalive: true
 706 | 
 707 | # Cassandra provides two out-of-the-box options for the RPC Server:
 708 | #
 709 | # sync
 710 | #   One thread per thrift connection. For a very large number of clients, memory
 711 | #   will be your limiting factor. On a 64 bit JVM, 180KB is the minimum stack size
 712 | #   per thread, and that will correspond to your use of virtual memory (but physical memory
 713 | #   may be limited depending on use of stack space).
 714 | #
 715 | # hsha
 716 | #   Stands for "half synchronous, half asynchronous." All thrift clients are handled
 717 | #   asynchronously using a small number of threads that does not vary with the amount
 718 | #   of thrift clients (and thus scales well to many clients). The rpc requests are still
 719 | #   synchronous (one thread per active request). If hsha is selected then it is essential
 720 | #   that rpc_max_threads is changed from the default value of unlimited.
 721 | #
 722 | # The default is sync because on Windows hsha is about 30% slower.  On Linux,
 723 | # sync/hsha performance is about the same, with hsha of course using less memory.
 724 | #
 725 | # Alternatively,  can provide your own RPC server by providing the fully-qualified class name
 726 | # of an o.a.c.t.TServerFactory that can create an instance of it.
 727 | rpc_server_type: sync
 728 | 
 729 | # Uncomment rpc_min|max_thread to set request pool size limits.
 730 | #
 731 | # Regardless of your choice of RPC server (see above), the number of maximum requests in the
 732 | # RPC thread pool dictates how many concurrent requests are possible (but if you are using the sync
 733 | # RPC server, it also dictates the number of clients that can be connected at all).
 734 | #
 735 | # The default is unlimited and thus provides no protection against clients overwhelming the server. You are
 736 | # encouraged to set a maximum that makes sense for you in production, but do keep in mind that
 737 | # rpc_max_threads represents the maximum number of client requests this server may execute concurrently.
 738 | #
 739 | # rpc_min_threads: 16
 740 | # rpc_max_threads: 2048
 741 | 
 742 | # uncomment to set socket buffer sizes on rpc connections
 743 | # rpc_send_buff_size_in_bytes:
 744 | # rpc_recv_buff_size_in_bytes:
 745 | 
 746 | # Uncomment to set socket buffer size for internode communication
 747 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 748 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 749 | # See also:
 750 | # /proc/sys/net/core/wmem_max
 751 | # /proc/sys/net/core/rmem_max
 752 | # /proc/sys/net/ipv4/tcp_wmem
 753 | # /proc/sys/net/ipv4/tcp_wmem
 754 | # and 'man tcp'
 755 | # internode_send_buff_size_in_bytes:
 756 | 
 757 | # Uncomment to set socket buffer size for internode communication
 758 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 759 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 760 | # internode_recv_buff_size_in_bytes:
 761 | 
 762 | # Frame size for thrift (maximum message length).
 763 | thrift_framed_transport_size_in_mb: 15
 764 | 
 765 | # Set to true to have Cassandra create a hard link to each sstable
 766 | # flushed or streamed locally in a backups/ subdirectory of the
 767 | # keyspace data.  Removing these links is the operator's
 768 | # responsibility.
 769 | incremental_backups: false
 770 | 
 771 | # Whether or not to take a snapshot before each compaction.  Be
 772 | # careful using this option, since Cassandra won't clean up the
 773 | # snapshots for you.  Mostly useful if you're paranoid when there
 774 | # is a data format change.
 775 | snapshot_before_compaction: false
 776 | 
 777 | # Whether or not a snapshot is taken of the data before keyspace truncation
 778 | # or dropping of column families. The STRONGLY advised default of true 
 779 | # should be used to provide data safety. If you set this flag to false, you will
 780 | # lose data on truncation or drop.
 781 | auto_snapshot: true
 782 | 
 783 | # Granularity of the collation index of rows within a partition.
 784 | # Increase if your rows are large, or if you have a very large
 785 | # number of rows per partition.  The competing goals are these:
 786 | #
 787 | # - a smaller granularity means more index entries are generated
 788 | #   and looking up rows withing the partition by collation column
 789 | #   is faster
 790 | # - but, Cassandra will keep the collation index in memory for hot
 791 | #   rows (as part of the key cache), so a larger granularity means
 792 | #   you can cache more hot rows
 793 | column_index_size_in_kb: 64
 794 | 
 795 | # Per sstable indexed key cache entries (the collation index in memory
 796 | # mentioned above) exceeding this size will not be held on heap.
 797 | # This means that only partition information is held on heap and the
 798 | # index entries are read from disk.
 799 | #
 800 | # Note that this size refers to the size of the
 801 | # serialized index information and not the size of the partition.
 802 | column_index_cache_size_in_kb: 2
 803 | 
 804 | # Number of simultaneous compactions to allow, NOT including
 805 | # validation "compactions" for anti-entropy repair.  Simultaneous
 806 | # compactions can help preserve read performance in a mixed read/write
 807 | # workload, by mitigating the tendency of small sstables to accumulate
 808 | # during a single long running compactions. The default is usually
 809 | # fine and if you experience problems with compaction running too
 810 | # slowly or too fast, you should look at
 811 | # compaction_throughput_mb_per_sec first.
 812 | #
 813 | # concurrent_compactors defaults to the smaller of (number of disks,
 814 | # number of cores), with a minimum of 2 and a maximum of 8.
 815 | # 
 816 | # If your data directories are backed by SSD, you should increase this
 817 | # to the number of cores.
 818 | #concurrent_compactors: 1
 819 | 
 820 | # Throttles compaction to the given total throughput across the entire
 821 | # system. The faster you insert data, the faster you need to compact in
 822 | # order to keep the sstable count down, but in general, setting this to
 823 | # 16 to 32 times the rate you are inserting data is more than sufficient.
 824 | # Setting this to 0 disables throttling. Note that this account for all types
 825 | # of compaction, including validation compaction.
 826 | compaction_throughput_mb_per_sec: 16
 827 | 
 828 | # When compacting, the replacement sstable(s) can be opened before they
 829 | # are completely written, and used in place of the prior sstables for
 830 | # any range that has been written. This helps to smoothly transfer reads 
 831 | # between the sstables, reducing page cache churn and keeping hot rows hot
 832 | sstable_preemptive_open_interval_in_mb: 50
 833 | 
 834 | # Throttles all outbound streaming file transfers on this node to the
 835 | # given total throughput in Mbps. This is necessary because Cassandra does
 836 | # mostly sequential IO when streaming data during bootstrap or repair, which
 837 | # can lead to saturating the network connection and degrading rpc performance.
 838 | # When unset, the default is 200 Mbps or 25 MB/s.
 839 | # stream_throughput_outbound_megabits_per_sec: 200
 840 | 
 841 | # Throttles all streaming file transfer between the datacenters,
 842 | # this setting allows users to throttle inter dc stream throughput in addition
 843 | # to throttling all network stream traffic as configured with
 844 | # stream_throughput_outbound_megabits_per_sec
 845 | # When unset, the default is 200 Mbps or 25 MB/s
 846 | # inter_dc_stream_throughput_outbound_megabits_per_sec: 200
 847 | 
 848 | # How long the coordinator should wait for read operations to complete
 849 | read_request_timeout_in_ms: 5000
 850 | # How long the coordinator should wait for seq or index scans to complete
 851 | range_request_timeout_in_ms: 10000
 852 | # How long the coordinator should wait for writes to complete
 853 | write_request_timeout_in_ms: 2000
 854 | # How long the coordinator should wait for counter writes to complete
 855 | counter_write_request_timeout_in_ms: 5000
 856 | # How long a coordinator should continue to retry a CAS operation
 857 | # that contends with other proposals for the same row
 858 | cas_contention_timeout_in_ms: 1000
 859 | # How long the coordinator should wait for truncates to complete
 860 | # (This can be much longer, because unless auto_snapshot is disabled
 861 | # we need to flush first so we can snapshot before removing the data.)
 862 | truncate_request_timeout_in_ms: 60000
 863 | # The default timeout for other, miscellaneous operations
 864 | request_timeout_in_ms: 10000
 865 | 
 866 | # How long before a node logs slow queries. Select queries that take longer than
 867 | # this timeout to execute, will generate an aggregated log message, so that slow queries
 868 | # can be identified. Set this value to zero to disable slow query logging.
 869 | slow_query_log_timeout_in_ms: 500
 870 | 
 871 | # Enable operation timeout information exchange between nodes to accurately
 872 | # measure request timeouts.  If disabled, replicas will assume that requests
 873 | # were forwarded to them instantly by the coordinator, which means that
 874 | # under overload conditions we will waste that much extra time processing 
 875 | # already-timed-out requests.
 876 | #
 877 | # Warning: before enabling this property make sure to ntp is installed
 878 | # and the times are synchronized between the nodes.
 879 | cross_node_timeout: false
 880 | 
 881 | # Set keep-alive period for streaming
 882 | # This node will send a keep-alive message periodically with this period.
 883 | # If the node does not receive a keep-alive message from the peer for
 884 | # 2 keep-alive cycles the stream session times out and fail
 885 | # Default value is 300s (5 minutes), which means stalled stream
 886 | # times out in 10 minutes by default
 887 | # streaming_keep_alive_period_in_secs: 300
 888 | 
 889 | # phi value that must be reached for a host to be marked down.
 890 | # most users should never need to adjust this.
 891 | # phi_convict_threshold: 8
 892 | 
 893 | # endpoint_snitch -- Set this to a class that implements
 894 | # IEndpointSnitch.  The snitch has two functions:
 895 | #
 896 | # - it teaches Cassandra enough about your network topology to route
 897 | #   requests efficiently
 898 | # - it allows Cassandra to spread replicas around your cluster to avoid
 899 | #   correlated failures. It does this by grouping machines into
 900 | #   "datacenters" and "racks."  Cassandra will do its best not to have
 901 | #   more than one replica on the same "rack" (which may not actually
 902 | #   be a physical location)
 903 | #
 904 | # CASSANDRA WILL NOT ALLOW YOU TO SWITCH TO AN INCOMPATIBLE SNITCH
 905 | # ONCE DATA IS INSERTED INTO THE CLUSTER.  This would cause data loss.
 906 | # This means that if you start with the default SimpleSnitch, which
 907 | # locates every node on "rack1" in "datacenter1", your only options
 908 | # if you need to add another datacenter are GossipingPropertyFileSnitch
 909 | # (and the older PFS).  From there, if you want to migrate to an
 910 | # incompatible snitch like Ec2Snitch you can do it by adding new nodes
 911 | # under Ec2Snitch (which will locate them in a new "datacenter") and
 912 | # decommissioning the old ones.
 913 | #
 914 | # Out of the box, Cassandra provides:
 915 | #
 916 | # SimpleSnitch:
 917 | #    Treats Strategy order as proximity. This can improve cache
 918 | #    locality when disabling read repair.  Only appropriate for
 919 | #    single-datacenter deployments.
 920 | #
 921 | # GossipingPropertyFileSnitch
 922 | #    This should be your go-to snitch for production use.  The rack
 923 | #    and datacenter for the local node are defined in
 924 | #    cassandra-rackdc.properties and propagated to other nodes via
 925 | #    gossip.  If cassandra-topology.properties exists, it is used as a
 926 | #    fallback, allowing migration from the PropertyFileSnitch.
 927 | #
 928 | # PropertyFileSnitch:
 929 | #    Proximity is determined by rack and data center, which are
 930 | #    explicitly configured in cassandra-topology.properties.
 931 | #
 932 | # Ec2Snitch:
 933 | #    Appropriate for EC2 deployments in a single Region. Loads Region
 934 | #    and Availability Zone information from the EC2 API. The Region is
 935 | #    treated as the datacenter, and the Availability Zone as the rack.
 936 | #    Only private IPs are used, so this will not work across multiple
 937 | #    Regions.
 938 | #
 939 | # Ec2MultiRegionSnitch:
 940 | #    Uses public IPs as broadcast_address to allow cross-region
 941 | #    connectivity.  (Thus, you should set seed addresses to the public
 942 | #    IP as well.) You will need to open the storage_port or
 943 | #    ssl_storage_port on the public IP firewall.  (For intra-Region
 944 | #    traffic, Cassandra will switch to the private IP after
 945 | #    establishing a connection.)
 946 | #
 947 | # RackInferringSnitch:
 948 | #    Proximity is determined by rack and data center, which are
 949 | #    assumed to correspond to the 3rd and 2nd octet of each node's IP
 950 | #    address, respectively.  Unless this happens to match your
 951 | #    deployment conventions, this is best used as an example of
 952 | #    writing a custom Snitch class and is provided in that spirit.
 953 | #
 954 | # You can use a custom Snitch by setting this to the full class name
 955 | # of the snitch, which will be assumed to be on your classpath.
 956 | endpoint_snitch: SimpleSnitch
 957 | 
 958 | # controls how often to perform the more expensive part of host score
 959 | # calculation
 960 | dynamic_snitch_update_interval_in_ms: 100 
 961 | # controls how often to reset all host scores, allowing a bad host to
 962 | # possibly recover
 963 | dynamic_snitch_reset_interval_in_ms: 600000
 964 | # if set greater than zero and read_repair_chance is < 1.0, this will allow
 965 | # 'pinning' of replicas to hosts in order to increase cache capacity.
 966 | # The badness threshold will control how much worse the pinned host has to be
 967 | # before the dynamic snitch will prefer other replicas over it.  This is
 968 | # expressed as a double which represents a percentage.  Thus, a value of
 969 | # 0.2 means Cassandra would continue to prefer the static snitch values
 970 | # until the pinned host was 20% worse than the fastest.
 971 | dynamic_snitch_badness_threshold: 0.1
 972 | 
 973 | # request_scheduler -- Set this to a class that implements
 974 | # RequestScheduler, which will schedule incoming client requests
 975 | # according to the specific policy. This is useful for multi-tenancy
 976 | # with a single Cassandra cluster.
 977 | # NOTE: This is specifically for requests from the client and does
 978 | # not affect inter node communication.
 979 | # org.apache.cassandra.scheduler.NoScheduler - No scheduling takes place
 980 | # org.apache.cassandra.scheduler.RoundRobinScheduler - Round robin of
 981 | # client requests to a node with a separate queue for each
 982 | # request_scheduler_id. The scheduler is further customized by
 983 | # request_scheduler_options as described below.
 984 | request_scheduler: org.apache.cassandra.scheduler.NoScheduler
 985 | 
 986 | # Scheduler Options vary based on the type of scheduler
 987 | #
 988 | # NoScheduler
 989 | #   Has no options
 990 | #
 991 | # RoundRobin
 992 | #   throttle_limit
 993 | #     The throttle_limit is the number of in-flight
 994 | #     requests per client.  Requests beyond 
 995 | #     that limit are queued up until
 996 | #     running requests can complete.
 997 | #     The value of 80 here is twice the number of
 998 | #     concurrent_reads + concurrent_writes.
 999 | #   default_weight
1000 | #     default_weight is optional and allows for
1001 | #     overriding the default which is 1.
1002 | #   weights
1003 | #     Weights are optional and will default to 1 or the
1004 | #     overridden default_weight. The weight translates into how
1005 | #     many requests are handled during each turn of the
1006 | #     RoundRobin, based on the scheduler id.
1007 | #
1008 | # request_scheduler_options:
1009 | #    throttle_limit: 80
1010 | #    default_weight: 5
1011 | #    weights:
1012 | #      Keyspace1: 1
1013 | #      Keyspace2: 5
1014 | 
1015 | # request_scheduler_id -- An identifier based on which to perform
1016 | # the request scheduling. Currently the only valid option is keyspace.
1017 | # request_scheduler_id: keyspace
1018 | 
1019 | # Enable or disable inter-node encryption
1020 | # JVM defaults for supported SSL socket protocols and cipher suites can
1021 | # be replaced using custom encryption options. This is not recommended
1022 | # unless you have policies in place that dictate certain settings, or
1023 | # need to disable vulnerable ciphers or protocols in case the JVM cannot
1024 | # be updated.
1025 | # FIPS compliant settings can be configured at JVM level and should not
1026 | # involve changing encryption settings here:
1027 | # https://docs.oracle.com/javase/8/docs/technotes/guides/security/jsse/FIPS.html
1028 | # *NOTE* No custom encryption options are enabled at the moment
1029 | # The available internode options are : all, none, dc, rack
1030 | #
1031 | # If set to dc cassandra will encrypt the traffic between the DCs
1032 | # If set to rack cassandra will encrypt the traffic between the racks
1033 | #
1034 | # The passwords used in these options must match the passwords used when generating
1035 | # the keystore and truststore.  For instructions on generating these files, see:
1036 | # http://download.oracle.com/javase/6/docs/technotes/guides/security/jsse/JSSERefGuide.html#CreateKeystore
1037 | #
1038 | server_encryption_options:
1039 |     internode_encryption: none
1040 |     keystore: conf/.keystore
1041 |     keystore_password: cassandra
1042 |     truststore: conf/.truststore
1043 |     truststore_password: cassandra
1044 |     # More advanced defaults below:
1045 |     # protocol: TLS
1046 |     # algorithm: SunX509
1047 |     # store_type: JKS
1048 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1049 |     # require_client_auth: false
1050 |     # require_endpoint_verification: false
1051 | 
1052 | # enable or disable client/server encryption.
1053 | client_encryption_options:
1054 |     enabled: false
1055 |     # If enabled and optional is set to true encrypted and unencrypted connections are handled.
1056 |     optional: false
1057 |     keystore: conf/.keystore
1058 |     keystore_password: cassandra
1059 |     # require_client_auth: false
1060 |     # Set trustore and truststore_password if require_client_auth is true
1061 |     # truststore: conf/.truststore
1062 |     # truststore_password: cassandra
1063 |     # More advanced defaults below:
1064 |     # protocol: TLS
1065 |     # algorithm: SunX509
1066 |     # store_type: JKS
1067 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1068 | 
1069 | # internode_compression controls whether traffic between nodes is
1070 | # compressed.
1071 | # Can be:
1072 | #
1073 | # all
1074 | #   all traffic is compressed
1075 | #
1076 | # dc
1077 | #   traffic between different datacenters is compressed
1078 | #
1079 | # none
1080 | #   nothing is compressed.
1081 | internode_compression: dc
1082 | 
1083 | # Enable or disable tcp_nodelay for inter-dc communication.
1084 | # Disabling it will result in larger (but fewer) network packets being sent,
1085 | # reducing overhead from the TCP protocol itself, at the cost of increasing
1086 | # latency if you block for cross-datacenter responses.
1087 | inter_dc_tcp_nodelay: false
1088 | 
1089 | # TTL for different trace types used during logging of the repair process.
1090 | tracetype_query_ttl: 86400
1091 | tracetype_repair_ttl: 604800
1092 | 
1093 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1094 | # This threshold can be adjusted to minimize logging if necessary
1095 | # gc_log_threshold_in_ms: 200
1096 | 
1097 | # If unset, all GC Pauses greater than gc_log_threshold_in_ms will log at
1098 | # INFO level
1099 | # UDFs (user defined functions) are disabled by default.
1100 | # As of Cassandra 3.0 there is a sandbox in place that should prevent execution of evil code.
1101 | enable_user_defined_functions: false
1102 | 
1103 | # Enables scripted UDFs (JavaScript UDFs).
1104 | # Java UDFs are always enabled, if enable_user_defined_functions is true.
1105 | # Enable this option to be able to use UDFs with "language javascript" or any custom JSR-223 provider.
1106 | # This option has no effect, if enable_user_defined_functions is false.
1107 | enable_scripted_user_defined_functions: false
1108 | 
1109 | # The default Windows kernel timer and scheduling resolution is 15.6ms for power conservation.
1110 | # Lowering this value on Windows can provide much tighter latency and better throughput, however
1111 | # some virtualized environments may see a negative performance impact from changing this setting
1112 | # below their system default. The sysinternals 'clockres' tool can confirm your system's default
1113 | # setting.
1114 | windows_timer_interval: 1
1115 | 
1116 | 
1117 | # Enables encrypting data at-rest (on disk). Different key providers can be plugged in, but the default reads from
1118 | # a JCE-style keystore. A single keystore can hold multiple keys, but the one referenced by
1119 | # the "key_alias" is the only key that will be used for encrypt opertaions; previously used keys
1120 | # can still (and should!) be in the keystore and will be used on decrypt operations
1121 | # (to handle the case of key rotation).
1122 | #
1123 | # It is strongly recommended to download and install Java Cryptography Extension (JCE)
1124 | # Unlimited Strength Jurisdiction Policy Files for your version of the JDK.
1125 | # (current link: http://www.oracle.com/technetwork/java/javase/downloads/jce8-download-2133166.html)
1126 | #
1127 | # Currently, only the following file types are supported for transparent data encryption, although
1128 | # more are coming in future cassandra releases: commitlog, hints
1129 | transparent_data_encryption_options:
1130 |     enabled: false
1131 |     chunk_length_kb: 64
1132 |     cipher: AES/CBC/PKCS5Padding
1133 |     key_alias: testing:1
1134 |     # CBC IV length for AES needs to be 16 bytes (which is also the default size)
1135 |     # iv_length: 16
1136 |     key_provider: 
1137 |       - class_name: org.apache.cassandra.security.JKSKeyProvider
1138 |         parameters: 
1139 |           - keystore: conf/.keystore
1140 |             keystore_password: cassandra
1141 |             store_type: JCEKS
1142 |             key_password: cassandra
1143 | 
1144 | 
1145 | #####################
1146 | # SAFETY THRESHOLDS #
1147 | #####################
1148 | 
1149 | # When executing a scan, within or across a partition, we need to keep the
1150 | # tombstones seen in memory so we can return them to the coordinator, which
1151 | # will use them to make sure other replicas also know about the deleted rows.
1152 | # With workloads that generate a lot of tombstones, this can cause performance
1153 | # problems and even exaust the server heap.
1154 | # (http://www.datastax.com/dev/blog/cassandra-anti-patterns-queues-and-queue-like-datasets)
1155 | # Adjust the thresholds here if you understand the dangers and want to
1156 | # scan more tombstones anyway.  These thresholds may also be adjusted at runtime
1157 | # using the StorageService mbean.
1158 | tombstone_warn_threshold: 1000
1159 | tombstone_failure_threshold: 100000
1160 | 
1161 | # Log WARN on any multiple-partition batch size exceeding this value. 5kb per batch by default.
1162 | # Caution should be taken on increasing the size of this threshold as it can lead to node instability.
1163 | batch_size_warn_threshold_in_kb: 5
1164 | 
1165 | # Fail any multiple-partition batch exceeding this value. 50kb (10x warn threshold) by default.
1166 | batch_size_fail_threshold_in_kb: 50
1167 | 
1168 | # Log WARN on any batches not of type LOGGED than span across more partitions than this limit
1169 | unlogged_batch_across_partitions_warn_threshold: 10
1170 | 
1171 | # Log a warning when compacting partitions larger than this value
1172 | compaction_large_partition_warning_threshold_mb: 100
1173 | 
1174 | # GC Pauses greater than gc_warn_threshold_in_ms will be logged at WARN level
1175 | # Adjust the threshold based on your application throughput requirement
1176 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1177 | gc_warn_threshold_in_ms: 1000
1178 | 
1179 | # Maximum size of any value in SSTables. Safety measure to detect SSTable corruption
1180 | # early. Any value size larger than this threshold will result into marking an SSTable
1181 | # as corrupted.
1182 | # max_value_size_in_mb: 256
1183 | 
1184 | # Back-pressure settings #
1185 | # If enabled, the coordinator will apply the back-pressure strategy specified below to each mutation
1186 | # sent to replicas, with the aim of reducing pressure on overloaded replicas.
1187 | back_pressure_enabled: false
1188 | # The back-pressure strategy applied.
1189 | # The default implementation, RateBasedBackPressure, takes three arguments:
1190 | # high ratio, factor, and flow type, and uses the ratio between incoming mutation responses and outgoing mutation requests.
1191 | # If below high ratio, outgoing mutations are rate limited according to the incoming rate decreased by the given factor;
1192 | # if above high ratio, the rate limiting is increased by the given factor;
1193 | # such factor is usually best configured between 1 and 10, use larger values for a faster recovery
1194 | # at the expense of potentially more dropped mutations;
1195 | # the rate limiting is applied according to the flow type: if FAST, it's rate limited at the speed of the fastest replica,
1196 | # if SLOW at the speed of the slowest one.
1197 | # New strategies can be added. Implementors need to implement org.apache.cassandra.net.BackpressureStrategy and
1198 | # provide a public constructor accepting a Map<String, Object>.
1199 | back_pressure_strategy:
1200 |     - class_name: org.apache.cassandra.net.RateBasedBackPressure
1201 |       parameters:
1202 |         - high_ratio: 0.90
1203 |           factor: 5
1204 |           flow: FAST
1205 | 
1206 | # Coalescing Strategies #
1207 | # Coalescing multiples messages turns out to significantly boost message processing throughput (think doubling or more).
1208 | # On bare metal, the floor for packet processing throughput is high enough that many applications won't notice, but in
1209 | # virtualized environments, the point at which an application can be bound by network packet processing can be
1210 | # surprisingly low compared to the throughput of task processing that is possible inside a VM. It's not that bare metal
1211 | # doesn't benefit from coalescing messages, it's that the number of packets a bare metal network interface can process
1212 | # is sufficient for many applications such that no load starvation is experienced even without coalescing.
1213 | # There are other benefits to coalescing network messages that are harder to isolate with a simple metric like messages
1214 | # per second. By coalescing multiple tasks together, a network thread can process multiple messages for the cost of one
1215 | # trip to read from a socket, and all the task submission work can be done at the same time reducing context switching
1216 | # and increasing cache friendliness of network message processing.
1217 | # See CASSANDRA-8692 for details.
1218 | 
1219 | # Strategy to use for coalescing messages in OutboundTcpConnection.
1220 | # Can be fixed, movingaverage, timehorizon, disabled (default).
1221 | # You can also specify a subclass of CoalescingStrategies.CoalescingStrategy by name.
1222 | # otc_coalescing_strategy: DISABLED
1223 | 
1224 | # How many microseconds to wait for coalescing. For fixed strategy this is the amount of time after the first
1225 | # message is received before it will be sent with any accompanying messages. For moving average this is the
1226 | # maximum amount of time that will be waited as well as the interval at which messages must arrive on average
1227 | # for coalescing to be enabled.
1228 | # otc_coalescing_window_us: 200
1229 | 
1230 | # Do not try to coalesce messages if we already got that many messages. This should be more than 2 and less than 128.
1231 | # otc_coalescing_enough_coalesced_messages: 8
1232 | 
1233 | # How many milliseconds to wait between two expiration runs on the backlog (queue) of the OutboundTcpConnection.
1234 | # Expiration is done if messages are piling up in the backlog. Droppable messages are expired to free the memory
1235 | # taken by expired messages. The interval should be between 0 and 1000, and in most installations the default value
1236 | # will be appropriate. A smaller value could potentially expire messages slightly sooner at the expense of more CPU
1237 | # time and queue contention while iterating the backlog of messages.
1238 | # An interval of 0 disables any wait time, which is the behavior of former Cassandra versions.
1239 | #
1240 | # otc_backlog_expiration_interval_ms: 200
1241 | 


--------------------------------------------------------------------------------
/cassandra-cdc/config/reader-1.yml:
--------------------------------------------------------------------------------
 1 | cassandra:
 2 |   keyspace: custom
 3 |   table: movies_by_genre
 4 |   cdc_raw_directory: /tmp/cdc/cassandra-1/cdc_raw
 5 | kafka:
 6 |   topic: cdc-topic
 7 |   configuration:
 8 |     bootstrap.servers: 0.0.0.0:33028, 0.0.0.0:33029
 9 |     key.serializer: org.apache.kafka.common.serialization.StringSerializer
10 |     value.serializer: org.apache.kafka.common.serialization.StringSerializer
11 | 


--------------------------------------------------------------------------------
/cassandra-cdc/config/reader-2.yml:
--------------------------------------------------------------------------------
 1 | cassandra:
 2 |   keyspace: custom
 3 |   table: movies_by_genre
 4 |   cdc_raw_directory: /tmp/cdc/cassandra-2/cdc_raw
 5 | kafka:
 6 |   topic: cdc-topic
 7 |   configuration:
 8 |     bootstrap.servers: 0.0.0.0:33028, 0.0.0.0:33029
 9 |     key.serializer: org.apache.kafka.common.serialization.StringSerializer
10 |     value.serializer: org.apache.kafka.common.serialization.StringSerializer
11 | 


--------------------------------------------------------------------------------
/cassandra-cdc/docker/Dockerfile:
--------------------------------------------------------------------------------
1 | FROM cassandra:3.11.0
2 | COPY cassandra.yaml /etc/cassandra/cassandra.yaml
3 | CMD ["cassandra", "-f"]
4 | 


--------------------------------------------------------------------------------
/cassandra-cdc/docker/cassandra.yaml:
--------------------------------------------------------------------------------
   1 | # Cassandra storage config YAML
   2 | 
   3 | # NOTE:
   4 | #   See http://wiki.apache.org/cassandra/StorageConfiguration for
   5 | #   full explanations of configuration directives
   6 | # /NOTE
   7 | 
   8 | # The name of the cluster. This is mainly used to prevent machines in
   9 | # one logical cluster from joining another.
  10 | cluster_name: 'Test Cluster'
  11 | 
  12 | # This defines the number of tokens randomly assigned to this node on the ring
  13 | # The more tokens, relative to other nodes, the larger the proportion of data
  14 | # that this node will store. You probably want all nodes to have the same number
  15 | # of tokens assuming they have equal hardware capability.
  16 | #
  17 | # If you leave this unspecified, Cassandra will use the default of 1 token for legacy compatibility,
  18 | # and will use the initial_token as described below.
  19 | #
  20 | # Specifying initial_token will override this setting on the node's initial start,
  21 | # on subsequent starts, this setting will apply even if initial token is set.
  22 | #
  23 | # If you already have a cluster with 1 token per node, and wish to migrate to 
  24 | # multiple tokens per node, see http://wiki.apache.org/cassandra/Operations
  25 | num_tokens: 256
  26 | 
  27 | # Triggers automatic allocation of num_tokens tokens for this node. The allocation
  28 | # algorithm attempts to choose tokens in a way that optimizes replicated load over
  29 | # the nodes in the datacenter for the replication strategy used by the specified
  30 | # keyspace.
  31 | #
  32 | # The load assigned to each node will be close to proportional to its number of
  33 | # vnodes.
  34 | #
  35 | # Only supported with the Murmur3Partitioner.
  36 | # allocate_tokens_for_keyspace: KEYSPACE
  37 | 
  38 | # initial_token allows you to specify tokens manually.  While you can use it with
  39 | # vnodes (num_tokens > 1, above) -- in which case you should provide a 
  40 | # comma-separated list -- it's primarily used when adding nodes to legacy clusters 
  41 | # that do not have vnodes enabled.
  42 | # initial_token:
  43 | 
  44 | # See http://wiki.apache.org/cassandra/HintedHandoff
  45 | # May either be "true" or "false" to enable globally
  46 | hinted_handoff_enabled: true
  47 | 
  48 | # When hinted_handoff_enabled is true, a black list of data centers that will not
  49 | # perform hinted handoff
  50 | # hinted_handoff_disabled_datacenters:
  51 | #    - DC1
  52 | #    - DC2
  53 | 
  54 | # this defines the maximum amount of time a dead host will have hints
  55 | # generated.  After it has been dead this long, new hints for it will not be
  56 | # created until it has been seen alive and gone down again.
  57 | max_hint_window_in_ms: 10800000 # 3 hours
  58 | 
  59 | # Maximum throttle in KBs per second, per delivery thread.  This will be
  60 | # reduced proportionally to the number of nodes in the cluster.  (If there
  61 | # are two nodes in the cluster, each delivery thread will use the maximum
  62 | # rate; if there are three, each will throttle to half of the maximum,
  63 | # since we expect two nodes to be delivering hints simultaneously.)
  64 | hinted_handoff_throttle_in_kb: 1024
  65 | 
  66 | # Number of threads with which to deliver hints;
  67 | # Consider increasing this number when you have multi-dc deployments, since
  68 | # cross-dc handoff tends to be slower
  69 | max_hints_delivery_threads: 2
  70 | 
  71 | # Directory where Cassandra should store hints.
  72 | # If not set, the default directory is $CASSANDRA_HOME/data/hints.
  73 | # hints_directory: /var/lib/cassandra/hints
  74 | hints_directory: /var/lib/cassandra/hints
  75 | 
  76 | # How often hints should be flushed from the internal buffers to disk.
  77 | # Will *not* trigger fsync.
  78 | hints_flush_period_in_ms: 10000
  79 | 
  80 | # Maximum size for a single hints file, in megabytes.
  81 | max_hints_file_size_in_mb: 128
  82 | 
  83 | # Compression to apply to the hint files. If omitted, hints files
  84 | # will be written uncompressed. LZ4, Snappy, and Deflate compressors
  85 | # are supported.
  86 | #hints_compression:
  87 | #   - class_name: LZ4Compressor
  88 | #     parameters:
  89 | #         -
  90 | 
  91 | # Maximum throttle in KBs per second, total. This will be
  92 | # reduced proportionally to the number of nodes in the cluster.
  93 | batchlog_replay_throttle_in_kb: 1024
  94 | 
  95 | # Authentication backend, implementing IAuthenticator; used to identify users
  96 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthenticator,
  97 | # PasswordAuthenticator}.
  98 | #
  99 | # - AllowAllAuthenticator performs no checks - set it to disable authentication.
 100 | # - PasswordAuthenticator relies on username/password pairs to authenticate
 101 | #   users. It keeps usernames and hashed passwords in system_auth.roles table.
 102 | #   Please increase system_auth keyspace replication factor if you use this authenticator.
 103 | #   If using PasswordAuthenticator, CassandraRoleManager must also be used (see below)
 104 | authenticator: AllowAllAuthenticator
 105 | 
 106 | # Authorization backend, implementing IAuthorizer; used to limit access/provide permissions
 107 | # Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthorizer,
 108 | # CassandraAuthorizer}.
 109 | #
 110 | # - AllowAllAuthorizer allows any action to any user - set it to disable authorization.
 111 | # - CassandraAuthorizer stores permissions in system_auth.role_permissions table. Please
 112 | #   increase system_auth keyspace replication factor if you use this authorizer.
 113 | authorizer: AllowAllAuthorizer
 114 | 
 115 | # Part of the Authentication & Authorization backend, implementing IRoleManager; used
 116 | # to maintain grants and memberships between roles.
 117 | # Out of the box, Cassandra provides org.apache.cassandra.auth.CassandraRoleManager,
 118 | # which stores role information in the system_auth keyspace. Most functions of the
 119 | # IRoleManager require an authenticated login, so unless the configured IAuthenticator
 120 | # actually implements authentication, most of this functionality will be unavailable.
 121 | #
 122 | # - CassandraRoleManager stores role data in the system_auth keyspace. Please
 123 | #   increase system_auth keyspace replication factor if you use this role manager.
 124 | role_manager: CassandraRoleManager
 125 | 
 126 | # Validity period for roles cache (fetching granted roles can be an expensive
 127 | # operation depending on the role manager, CassandraRoleManager is one example)
 128 | # Granted roles are cached for authenticated sessions in AuthenticatedUser and
 129 | # after the period specified here, become eligible for (async) reload.
 130 | # Defaults to 2000, set to 0 to disable caching entirely.
 131 | # Will be disabled automatically for AllowAllAuthenticator.
 132 | roles_validity_in_ms: 2000
 133 | 
 134 | # Refresh interval for roles cache (if enabled).
 135 | # After this interval, cache entries become eligible for refresh. Upon next
 136 | # access, an async reload is scheduled and the old value returned until it
 137 | # completes. If roles_validity_in_ms is non-zero, then this must be
 138 | # also.
 139 | # Defaults to the same value as roles_validity_in_ms.
 140 | # roles_update_interval_in_ms: 2000
 141 | 
 142 | # Validity period for permissions cache (fetching permissions can be an
 143 | # expensive operation depending on the authorizer, CassandraAuthorizer is
 144 | # one example). Defaults to 2000, set to 0 to disable.
 145 | # Will be disabled automatically for AllowAllAuthorizer.
 146 | permissions_validity_in_ms: 2000
 147 | 
 148 | # Refresh interval for permissions cache (if enabled).
 149 | # After this interval, cache entries become eligible for refresh. Upon next
 150 | # access, an async reload is scheduled and the old value returned until it
 151 | # completes. If permissions_validity_in_ms is non-zero, then this must be
 152 | # also.
 153 | # Defaults to the same value as permissions_validity_in_ms.
 154 | # permissions_update_interval_in_ms: 2000
 155 | 
 156 | # Validity period for credentials cache. This cache is tightly coupled to
 157 | # the provided PasswordAuthenticator implementation of IAuthenticator. If
 158 | # another IAuthenticator implementation is configured, this cache will not
 159 | # be automatically used and so the following settings will have no effect.
 160 | # Please note, credentials are cached in their encrypted form, so while
 161 | # activating this cache may reduce the number of queries made to the
 162 | # underlying table, it may not  bring a significant reduction in the
 163 | # latency of individual authentication attempts.
 164 | # Defaults to 2000, set to 0 to disable credentials caching.
 165 | credentials_validity_in_ms: 2000
 166 | 
 167 | # Refresh interval for credentials cache (if enabled).
 168 | # After this interval, cache entries become eligible for refresh. Upon next
 169 | # access, an async reload is scheduled and the old value returned until it
 170 | # completes. If credentials_validity_in_ms is non-zero, then this must be
 171 | # also.
 172 | # Defaults to the same value as credentials_validity_in_ms.
 173 | # credentials_update_interval_in_ms: 2000
 174 | 
 175 | # The partitioner is responsible for distributing groups of rows (by
 176 | # partition key) across nodes in the cluster.  You should leave this
 177 | # alone for new clusters.  The partitioner can NOT be changed without
 178 | # reloading all data, so when upgrading you should set this to the
 179 | # same partitioner you were already using.
 180 | #
 181 | # Besides Murmur3Partitioner, partitioners included for backwards
 182 | # compatibility include RandomPartitioner, ByteOrderedPartitioner, and
 183 | # OrderPreservingPartitioner.
 184 | #
 185 | partitioner: org.apache.cassandra.dht.Murmur3Partitioner
 186 | 
 187 | # Directories where Cassandra should store data on disk.  Cassandra
 188 | # will spread data evenly across them, subject to the granularity of
 189 | # the configured compaction strategy.
 190 | # If not set, the default directory is $CASSANDRA_HOME/data/data.
 191 | # data_file_directories:
 192 | #     - /var/lib/cassandra/data
 193 | data_file_directories:
 194 |   - /var/lib/cassandra/data
 195 | 
 196 | # commit log.  when running on magnetic HDD, this should be a
 197 | # separate spindle than the data directories.
 198 | # If not set, the default directory is $CASSANDRA_HOME/data/commitlog.
 199 | # commitlog_directory: /var/lib/cassandra/commitlog
 200 | commitlog_directory: /var/lib/cassandra/commitlog
 201 | 
 202 | # Enable / disable CDC functionality on a per-node basis. This modifies the logic used
 203 | # for write path allocation rejection (standard: never reject. cdc: reject Mutation
 204 | # containing a CDC-enabled table if at space limit in cdc_raw_directory).
 205 | cdc_enabled: true
 206 | 
 207 | # CommitLogSegments are moved to this directory on flush if cdc_enabled: true and the
 208 | # segment contains mutations for a CDC-enabled table. This should be placed on a
 209 | # separate spindle than the data directories. If not set, the default directory is
 210 | # $CASSANDRA_HOME/data/cdc_raw.
 211 | # cdc_raw_directory: /var/lib/cassandra/cdc_raw
 212 | cdc_raw_directory: /var/lib/cassandra/cdc_raw
 213 | 
 214 | # Policy for data disk failures:
 215 | #
 216 | # die
 217 | #   shut down gossip and client transports and kill the JVM for any fs errors or
 218 | #   single-sstable errors, so the node can be replaced.
 219 | #
 220 | # stop_paranoid
 221 | #   shut down gossip and client transports even for single-sstable errors,
 222 | #   kill the JVM for errors during startup.
 223 | #
 224 | # stop
 225 | #   shut down gossip and client transports, leaving the node effectively dead, but
 226 | #   can still be inspected via JMX, kill the JVM for errors during startup.
 227 | #
 228 | # best_effort
 229 | #    stop using the failed disk and respond to requests based on
 230 | #    remaining available sstables.  This means you WILL see obsolete
 231 | #    data at CL.ONE!
 232 | #
 233 | # ignore
 234 | #    ignore fatal errors and let requests fail, as in pre-1.2 Cassandra
 235 | disk_failure_policy: stop
 236 | 
 237 | # Policy for commit disk failures:
 238 | #
 239 | # die
 240 | #   shut down gossip and Thrift and kill the JVM, so the node can be replaced.
 241 | #
 242 | # stop
 243 | #   shut down gossip and Thrift, leaving the node effectively dead, but
 244 | #   can still be inspected via JMX.
 245 | #
 246 | # stop_commit
 247 | #   shutdown the commit log, letting writes collect but
 248 | #   continuing to service reads, as in pre-2.0.5 Cassandra
 249 | #
 250 | # ignore
 251 | #   ignore fatal errors and let the batches fail
 252 | commit_failure_policy: stop
 253 | 
 254 | # Maximum size of the native protocol prepared statement cache
 255 | #
 256 | # Valid values are either "auto" (omitting the value) or a value greater 0.
 257 | #
 258 | # Note that specifying a too large value will result in long running GCs and possbily
 259 | # out-of-memory errors. Keep the value at a small fraction of the heap.
 260 | #
 261 | # If you constantly see "prepared statements discarded in the last minute because
 262 | # cache limit reached" messages, the first step is to investigate the root cause
 263 | # of these messages and check whether prepared statements are used correctly -
 264 | # i.e. use bind markers for variable parts.
 265 | #
 266 | # Do only change the default value, if you really have more prepared statements than
 267 | # fit in the cache. In most cases it is not neccessary to change this value.
 268 | # Constantly re-preparing statements is a performance penalty.
 269 | #
 270 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 271 | prepared_statements_cache_size_mb:
 272 | 
 273 | # Maximum size of the Thrift prepared statement cache
 274 | #
 275 | # If you do not use Thrift at all, it is safe to leave this value at "auto".
 276 | #
 277 | # See description of 'prepared_statements_cache_size_mb' above for more information.
 278 | #
 279 | # Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater
 280 | thrift_prepared_statements_cache_size_mb:
 281 | 
 282 | # Maximum size of the key cache in memory.
 283 | #
 284 | # Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at the
 285 | # minimum, sometimes more. The key cache is fairly tiny for the amount of
 286 | # time it saves, so it's worthwhile to use it at large numbers.
 287 | # The row cache saves even more time, but must contain the entire row,
 288 | # so it is extremely space-intensive. It's best to only use the
 289 | # row cache if you have hot rows or static rows.
 290 | #
 291 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 292 | #
 293 | # Default value is empty to make it "auto" (min(5% of Heap (in MB), 100MB)). Set to 0 to disable key cache.
 294 | key_cache_size_in_mb:
 295 | 
 296 | # Duration in seconds after which Cassandra should
 297 | # save the key cache. Caches are saved to saved_caches_directory as
 298 | # specified in this configuration file.
 299 | #
 300 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 301 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 302 | # has limited use.
 303 | #
 304 | # Default is 14400 or 4 hours.
 305 | key_cache_save_period: 14400
 306 | 
 307 | # Number of keys from the key cache to save
 308 | # Disabled by default, meaning all keys are going to be saved
 309 | # key_cache_keys_to_save: 100
 310 | 
 311 | # Row cache implementation class name. Available implementations:
 312 | #
 313 | # org.apache.cassandra.cache.OHCProvider
 314 | #   Fully off-heap row cache implementation (default).
 315 | #
 316 | # org.apache.cassandra.cache.SerializingCacheProvider
 317 | #   This is the row cache implementation availabile
 318 | #   in previous releases of Cassandra.
 319 | # row_cache_class_name: org.apache.cassandra.cache.OHCProvider
 320 | 
 321 | # Maximum size of the row cache in memory.
 322 | # Please note that OHC cache implementation requires some additional off-heap memory to manage
 323 | # the map structures and some in-flight memory during operations before/after cache entries can be
 324 | # accounted against the cache capacity. This overhead is usually small compared to the whole capacity.
 325 | # Do not specify more memory that the system can afford in the worst usual situation and leave some
 326 | # headroom for OS block level cache. Do never allow your system to swap.
 327 | #
 328 | # Default value is 0, to disable row caching.
 329 | row_cache_size_in_mb: 0
 330 | 
 331 | # Duration in seconds after which Cassandra should save the row cache.
 332 | # Caches are saved to saved_caches_directory as specified in this configuration file.
 333 | #
 334 | # Saved caches greatly improve cold-start speeds, and is relatively cheap in
 335 | # terms of I/O for the key cache. Row cache saving is much more expensive and
 336 | # has limited use.
 337 | #
 338 | # Default is 0 to disable saving the row cache.
 339 | row_cache_save_period: 0
 340 | 
 341 | # Number of keys from the row cache to save.
 342 | # Specify 0 (which is the default), meaning all keys are going to be saved
 343 | # row_cache_keys_to_save: 100
 344 | 
 345 | # Maximum size of the counter cache in memory.
 346 | #
 347 | # Counter cache helps to reduce counter locks' contention for hot counter cells.
 348 | # In case of RF = 1 a counter cache hit will cause Cassandra to skip the read before
 349 | # write entirely. With RF > 1 a counter cache hit will still help to reduce the duration
 350 | # of the lock hold, helping with hot counter cell updates, but will not allow skipping
 351 | # the read entirely. Only the local (clock, count) tuple of a counter cell is kept
 352 | # in memory, not the whole counter, so it's relatively cheap.
 353 | #
 354 | # NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.
 355 | #
 356 | # Default value is empty to make it "auto" (min(2.5% of Heap (in MB), 50MB)). Set to 0 to disable counter cache.
 357 | # NOTE: if you perform counter deletes and rely on low gcgs, you should disable the counter cache.
 358 | counter_cache_size_in_mb:
 359 | 
 360 | # Duration in seconds after which Cassandra should
 361 | # save the counter cache (keys only). Caches are saved to saved_caches_directory as
 362 | # specified in this configuration file.
 363 | #
 364 | # Default is 7200 or 2 hours.
 365 | counter_cache_save_period: 7200
 366 | 
 367 | # Number of keys from the counter cache to save
 368 | # Disabled by default, meaning all keys are going to be saved
 369 | # counter_cache_keys_to_save: 100
 370 | 
 371 | # saved caches
 372 | # If not set, the default directory is $CASSANDRA_HOME/data/saved_caches.
 373 | # saved_caches_directory: /var/lib/cassandra/saved_caches
 374 | 
 375 | # commitlog_sync may be either "periodic" or "batch." 
 376 | # 
 377 | # When in batch mode, Cassandra won't ack writes until the commit log
 378 | # has been fsynced to disk.  It will wait
 379 | # commitlog_sync_batch_window_in_ms milliseconds between fsyncs.
 380 | # This window should be kept short because the writer threads will
 381 | # be unable to do extra work while waiting.  (You may need to increase
 382 | # concurrent_writes for the same reason.)
 383 | #
 384 | # commitlog_sync: batch
 385 | # commitlog_sync_batch_window_in_ms: 2
 386 | #
 387 | # the other option is "periodic" where writes may be acked immediately
 388 | # and the CommitLog is simply synced every commitlog_sync_period_in_ms
 389 | # milliseconds. 
 390 | commitlog_sync: periodic
 391 | commitlog_sync_period_in_ms: 10000
 392 | 
 393 | # The size of the individual commitlog file segments.  A commitlog
 394 | # segment may be archived, deleted, or recycled once all the data
 395 | # in it (potentially from each columnfamily in the system) has been
 396 | # flushed to sstables.
 397 | #
 398 | # The default size is 32, which is almost always fine, but if you are
 399 | # archiving commitlog segments (see commitlog_archiving.properties),
 400 | # then you probably want a finer granularity of archiving; 8 or 16 MB
 401 | # is reasonable.
 402 | # Max mutation size is also configurable via max_mutation_size_in_kb setting in
 403 | # cassandra.yaml. The default is half the size commitlog_segment_size_in_mb * 1024.
 404 | #
 405 | # NOTE: If max_mutation_size_in_kb is set explicitly then commitlog_segment_size_in_mb must
 406 | # be set to at least twice the size of max_mutation_size_in_kb / 1024
 407 | #
 408 | commitlog_segment_size_in_mb: 1
 409 | 
 410 | # Compression to apply to the commit log. If omitted, the commit log
 411 | # will be written uncompressed.  LZ4, Snappy, and Deflate compressors
 412 | # are supported.
 413 | # commitlog_compression:
 414 | #   - class_name: LZ4Compressor
 415 | #     parameters:
 416 | #         -
 417 | 
 418 | # any class that implements the SeedProvider interface and has a
 419 | # constructor that takes a Map<String, String> of parameters will do.
 420 | seed_provider:
 421 |     # Addresses of hosts that are deemed contact points. 
 422 |     # Cassandra nodes use this list of hosts to find each other and learn
 423 |     # the topology of the ring.  You must change this if you are running
 424 |     # multiple nodes!
 425 |     - class_name: org.apache.cassandra.locator.SimpleSeedProvider
 426 |       parameters:
 427 |           # seeds is actually a comma-delimited list of addresses.
 428 |           # Ex: "<ip1>,<ip2>,<ip3>"
 429 |           - seeds: "127.0.0.1"
 430 | 
 431 | # For workloads with more data than can fit in memory, Cassandra's
 432 | # bottleneck will be reads that need to fetch data from
 433 | # disk. "concurrent_reads" should be set to (16 * number_of_drives) in
 434 | # order to allow the operations to enqueue low enough in the stack
 435 | # that the OS and drives can reorder them. Same applies to
 436 | # "concurrent_counter_writes", since counter writes read the current
 437 | # values before incrementing and writing them back.
 438 | #
 439 | # On the other hand, since writes are almost never IO bound, the ideal
 440 | # number of "concurrent_writes" is dependent on the number of cores in
 441 | # your system; (8 * number_of_cores) is a good rule of thumb.
 442 | concurrent_reads: 32
 443 | concurrent_writes: 32
 444 | concurrent_counter_writes: 32
 445 | 
 446 | # For materialized view writes, as there is a read involved, so this should
 447 | # be limited by the less of concurrent reads or concurrent writes.
 448 | concurrent_materialized_view_writes: 32
 449 | 
 450 | # Maximum memory to use for sstable chunk cache and buffer pooling.
 451 | # 32MB of this are reserved for pooling buffers, the rest is used as an
 452 | # cache that holds uncompressed sstable chunks.
 453 | # Defaults to the smaller of 1/4 of heap or 512MB. This pool is allocated off-heap,
 454 | # so is in addition to the memory allocated for heap. The cache also has on-heap
 455 | # overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size
 456 | # if the default 64k chunk size is used).
 457 | # Memory is only allocated when needed.
 458 | # file_cache_size_in_mb: 512
 459 | 
 460 | # Flag indicating whether to allocate on or off heap when the sstable buffer
 461 | # pool is exhausted, that is when it has exceeded the maximum memory
 462 | # file_cache_size_in_mb, beyond which it will not cache buffers but allocate on request.
 463 | 
 464 | # buffer_pool_use_heap_if_exhausted: true
 465 | 
 466 | # The strategy for optimizing disk read
 467 | # Possible values are:
 468 | # ssd (for solid state disks, the default)
 469 | # spinning (for spinning disks)
 470 | # disk_optimization_strategy: ssd
 471 | 
 472 | # Total permitted memory to use for memtables. Cassandra will stop
 473 | # accepting writes when the limit is exceeded until a flush completes,
 474 | # and will trigger a flush based on memtable_cleanup_threshold
 475 | # If omitted, Cassandra will set both to 1/4 the size of the heap.
 476 | # memtable_heap_space_in_mb: 2048
 477 | # memtable_offheap_space_in_mb: 2048
 478 | 
 479 | # memtable_cleanup_threshold is deprecated. The default calculation
 480 | # is the only reasonable choice. See the comments on  memtable_flush_writers
 481 | # for more information.
 482 | #
 483 | # Ratio of occupied non-flushing memtable size to total permitted size
 484 | # that will trigger a flush of the largest memtable. Larger mct will
 485 | # mean larger flushes and hence less compaction, but also less concurrent
 486 | # flush activity which can make it difficult to keep your disks fed
 487 | # under heavy write load.
 488 | #
 489 | # memtable_cleanup_threshold defaults to 1 / (memtable_flush_writers + 1)
 490 | # memtable_cleanup_threshold: 0.11
 491 | 
 492 | # Specify the way Cassandra allocates and manages memtable memory.
 493 | # Options are:
 494 | #
 495 | # heap_buffers
 496 | #   on heap nio buffers
 497 | #
 498 | # offheap_buffers
 499 | #   off heap (direct) nio buffers
 500 | #
 501 | # offheap_objects
 502 | #    off heap objects
 503 | memtable_allocation_type: heap_buffers
 504 | 
 505 | # Total space to use for commit logs on disk.
 506 | #
 507 | # If space gets above this value, Cassandra will flush every dirty CF
 508 | # in the oldest segment and remove it.  So a small total commitlog space
 509 | # will tend to cause more flush activity on less-active columnfamilies.
 510 | #
 511 | # The default value is the smaller of 8192, and 1/4 of the total space
 512 | # of the commitlog volume.
 513 | #
 514 | # commitlog_total_space_in_mb: 8192
 515 | commitlog_total_space_in_mb: 16
 516 | 
 517 | # This sets the number of memtable flush writer threads per disk
 518 | # as well as the total number of memtables that can be flushed concurrently.
 519 | # These are generally a combination of compute and IO bound.
 520 | #
 521 | # Memtable flushing is more CPU efficient than memtable ingest and a single thread
 522 | # can keep up with the ingest rate of a whole server on a single fast disk
 523 | # until it temporarily becomes IO bound under contention typically with compaction.
 524 | # At that point you need multiple flush threads. At some point in the future
 525 | # it may become CPU bound all the time.
 526 | #
 527 | # You can tell if flushing is falling behind using the MemtablePool.BlockedOnAllocation
 528 | # metric which should be 0, but will be non-zero if threads are blocked waiting on flushing
 529 | # to free memory.
 530 | #
 531 | # memtable_flush_writers defaults to two for a single data directory.
 532 | # This means that two  memtables can be flushed concurrently to the single data directory.
 533 | # If you have multiple data directories the default is one memtable flushing at a time
 534 | # but the flush will use a thread per data directory so you will get two or more writers.
 535 | #
 536 | # Two is generally enough to flush on a fast disk [array] mounted as a single data directory.
 537 | # Adding more flush writers will result in smaller more frequent flushes that introduce more
 538 | # compaction overhead.
 539 | #
 540 | # There is a direct tradeoff between number of memtables that can be flushed concurrently
 541 | # and flush size and frequency. More is not better you just need enough flush writers
 542 | # to never stall waiting for flushing to free memory.
 543 | #
 544 | #memtable_flush_writers: 2
 545 | 
 546 | # Total space to use for change-data-capture logs on disk.
 547 | #
 548 | # If space gets above this value, Cassandra will throw WriteTimeoutException
 549 | # on Mutations including tables with CDC enabled. A CDCCompactor is responsible
 550 | # for parsing the raw CDC logs and deleting them when parsing is completed.
 551 | #
 552 | # The default value is the min of 4096 mb and 1/8th of the total space
 553 | # of the drive where cdc_raw_directory resides.
 554 | # cdc_total_space_in_mb: 4096
 555 | cdc_total_space_in_mb: 4096
 556 | 
 557 | # When we hit our cdc_raw limit and the CDCCompactor is either running behind
 558 | # or experiencing backpressure, we check at the following interval to see if any
 559 | # new space for cdc-tracked tables has been made available. Default to 250ms
 560 | # cdc_free_space_check_interval_ms: 250
 561 | cdc_free_space_check_interval_ms: 250
 562 | 
 563 | # A fixed memory pool size in MB for for SSTable index summaries. If left
 564 | # empty, this will default to 5% of the heap size. If the memory usage of
 565 | # all index summaries exceeds this limit, SSTables with low read rates will
 566 | # shrink their index summaries in order to meet this limit.  However, this
 567 | # is a best-effort process. In extreme conditions Cassandra may need to use
 568 | # more than this amount of memory.
 569 | index_summary_capacity_in_mb:
 570 | 
 571 | # How frequently index summaries should be resampled.  This is done
 572 | # periodically to redistribute memory from the fixed-size pool to sstables
 573 | # proportional their recent read rates.  Setting to -1 will disable this
 574 | # process, leaving existing index summaries at their current sampling level.
 575 | index_summary_resize_interval_in_minutes: 60
 576 | 
 577 | # Whether to, when doing sequential writing, fsync() at intervals in
 578 | # order to force the operating system to flush the dirty
 579 | # buffers. Enable this to avoid sudden dirty buffer flushing from
 580 | # impacting read latencies. Almost always a good idea on SSDs; not
 581 | # necessarily on platters.
 582 | trickle_fsync: false
 583 | trickle_fsync_interval_in_kb: 10240
 584 | 
 585 | # TCP port, for commands and data
 586 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 587 | storage_port: 7000
 588 | 
 589 | # SSL port, for encrypted communication.  Unused unless enabled in
 590 | # encryption_options
 591 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 592 | ssl_storage_port: 7001
 593 | 
 594 | # Address or interface to bind to and tell other Cassandra nodes to connect to.
 595 | # You _must_ change this if you want multiple nodes to be able to communicate!
 596 | #
 597 | # Set listen_address OR listen_interface, not both.
 598 | #
 599 | # Leaving it blank leaves it up to InetAddress.getLocalHost(). This
 600 | # will always do the Right Thing _if_ the node is properly configured
 601 | # (hostname, name resolution, etc), and the Right Thing is to use the
 602 | # address associated with the hostname (it might not be).
 603 | #
 604 | # Setting listen_address to 0.0.0.0 is always wrong.
 605 | #
 606 | listen_address: localhost
 607 | 
 608 | # Set listen_address OR listen_interface, not both. Interfaces must correspond
 609 | # to a single address, IP aliasing is not supported.
 610 | # listen_interface: eth0
 611 | 
 612 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 613 | # you can specify which should be chosen using listen_interface_prefer_ipv6. If false the first ipv4
 614 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 615 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 616 | # listen_interface_prefer_ipv6: false
 617 | 
 618 | # Address to broadcast to other Cassandra nodes
 619 | # Leaving this blank will set it to the same value as listen_address
 620 | # broadcast_address: 1.2.3.4
 621 | 
 622 | # When using multiple physical network interfaces, set this
 623 | # to true to listen on broadcast_address in addition to
 624 | # the listen_address, allowing nodes to communicate in both
 625 | # interfaces.
 626 | # Ignore this property if the network configuration automatically
 627 | # routes  between the public and private networks such as EC2.
 628 | # listen_on_broadcast_address: false
 629 | 
 630 | # Internode authentication backend, implementing IInternodeAuthenticator;
 631 | # used to allow/disallow connections from peer nodes.
 632 | # internode_authenticator: org.apache.cassandra.auth.AllowAllInternodeAuthenticator
 633 | 
 634 | # Whether to start the native transport server.
 635 | # Please note that the address on which the native transport is bound is the
 636 | # same as the rpc_address. The port however is different and specified below.
 637 | start_native_transport: true
 638 | # port for the CQL native transport to listen for clients on
 639 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 640 | native_transport_port: 9042
 641 | # Enabling native transport encryption in client_encryption_options allows you to either use
 642 | # encryption for the standard port or to use a dedicated, additional port along with the unencrypted
 643 | # standard native_transport_port.
 644 | # Enabling client encryption and keeping native_transport_port_ssl disabled will use encryption
 645 | # for native_transport_port. Setting native_transport_port_ssl to a different value
 646 | # from native_transport_port will use encryption for native_transport_port_ssl while
 647 | # keeping native_transport_port unencrypted.
 648 | # native_transport_port_ssl: 9142
 649 | # The maximum threads for handling requests when the native transport is used.
 650 | # This is similar to rpc_max_threads though the default differs slightly (and
 651 | # there is no native_transport_min_threads, idle threads will always be stopped
 652 | # after 30 seconds).
 653 | # native_transport_max_threads: 128
 654 | #
 655 | # The maximum size of allowed frame. Frame (requests) larger than this will
 656 | # be rejected as invalid. The default is 256MB. If you're changing this parameter,
 657 | # you may want to adjust max_value_size_in_mb accordingly.
 658 | # native_transport_max_frame_size_in_mb: 256
 659 | 
 660 | # The maximum number of concurrent client connections.
 661 | # The default is -1, which means unlimited.
 662 | # native_transport_max_concurrent_connections: -1
 663 | 
 664 | # The maximum number of concurrent client connections per source ip.
 665 | # The default is -1, which means unlimited.
 666 | # native_transport_max_concurrent_connections_per_ip: -1
 667 | 
 668 | # Whether to start the thrift rpc server.
 669 | start_rpc: false
 670 | 
 671 | # The address or interface to bind the Thrift RPC service and native transport
 672 | # server to.
 673 | #
 674 | # Set rpc_address OR rpc_interface, not both.
 675 | #
 676 | # Leaving rpc_address blank has the same effect as on listen_address
 677 | # (i.e. it will be based on the configured hostname of the node).
 678 | #
 679 | # Note that unlike listen_address, you can specify 0.0.0.0, but you must also
 680 | # set broadcast_rpc_address to a value other than 0.0.0.0.
 681 | #
 682 | # For security reasons, you should not expose this port to the internet.  Firewall it if needed.
 683 | rpc_address: localhost
 684 | 
 685 | # Set rpc_address OR rpc_interface, not both. Interfaces must correspond
 686 | # to a single address, IP aliasing is not supported.
 687 | # rpc_interface: eth1
 688 | 
 689 | # If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address
 690 | # you can specify which should be chosen using rpc_interface_prefer_ipv6. If false the first ipv4
 691 | # address will be used. If true the first ipv6 address will be used. Defaults to false preferring
 692 | # ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.
 693 | # rpc_interface_prefer_ipv6: false
 694 | 
 695 | # port for Thrift to listen for clients on
 696 | rpc_port: 9160
 697 | 
 698 | # RPC address to broadcast to drivers and other Cassandra nodes. This cannot
 699 | # be set to 0.0.0.0. If left blank, this will be set to the value of
 700 | # rpc_address. If rpc_address is set to 0.0.0.0, broadcast_rpc_address must
 701 | # be set.
 702 | # broadcast_rpc_address: 1.2.3.4
 703 | 
 704 | # enable or disable keepalive on rpc/native connections
 705 | rpc_keepalive: true
 706 | 
 707 | # Cassandra provides two out-of-the-box options for the RPC Server:
 708 | #
 709 | # sync
 710 | #   One thread per thrift connection. For a very large number of clients, memory
 711 | #   will be your limiting factor. On a 64 bit JVM, 180KB is the minimum stack size
 712 | #   per thread, and that will correspond to your use of virtual memory (but physical memory
 713 | #   may be limited depending on use of stack space).
 714 | #
 715 | # hsha
 716 | #   Stands for "half synchronous, half asynchronous." All thrift clients are handled
 717 | #   asynchronously using a small number of threads that does not vary with the amount
 718 | #   of thrift clients (and thus scales well to many clients). The rpc requests are still
 719 | #   synchronous (one thread per active request). If hsha is selected then it is essential
 720 | #   that rpc_max_threads is changed from the default value of unlimited.
 721 | #
 722 | # The default is sync because on Windows hsha is about 30% slower.  On Linux,
 723 | # sync/hsha performance is about the same, with hsha of course using less memory.
 724 | #
 725 | # Alternatively,  can provide your own RPC server by providing the fully-qualified class name
 726 | # of an o.a.c.t.TServerFactory that can create an instance of it.
 727 | rpc_server_type: sync
 728 | 
 729 | # Uncomment rpc_min|max_thread to set request pool size limits.
 730 | #
 731 | # Regardless of your choice of RPC server (see above), the number of maximum requests in the
 732 | # RPC thread pool dictates how many concurrent requests are possible (but if you are using the sync
 733 | # RPC server, it also dictates the number of clients that can be connected at all).
 734 | #
 735 | # The default is unlimited and thus provides no protection against clients overwhelming the server. You are
 736 | # encouraged to set a maximum that makes sense for you in production, but do keep in mind that
 737 | # rpc_max_threads represents the maximum number of client requests this server may execute concurrently.
 738 | #
 739 | # rpc_min_threads: 16
 740 | # rpc_max_threads: 2048
 741 | 
 742 | # uncomment to set socket buffer sizes on rpc connections
 743 | # rpc_send_buff_size_in_bytes:
 744 | # rpc_recv_buff_size_in_bytes:
 745 | 
 746 | # Uncomment to set socket buffer size for internode communication
 747 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 748 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 749 | # See also:
 750 | # /proc/sys/net/core/wmem_max
 751 | # /proc/sys/net/core/rmem_max
 752 | # /proc/sys/net/ipv4/tcp_wmem
 753 | # /proc/sys/net/ipv4/tcp_wmem
 754 | # and 'man tcp'
 755 | # internode_send_buff_size_in_bytes:
 756 | 
 757 | # Uncomment to set socket buffer size for internode communication
 758 | # Note that when setting this, the buffer size is limited by net.core.wmem_max
 759 | # and when not setting it it is defined by net.ipv4.tcp_wmem
 760 | # internode_recv_buff_size_in_bytes:
 761 | 
 762 | # Frame size for thrift (maximum message length).
 763 | thrift_framed_transport_size_in_mb: 15
 764 | 
 765 | # Set to true to have Cassandra create a hard link to each sstable
 766 | # flushed or streamed locally in a backups/ subdirectory of the
 767 | # keyspace data.  Removing these links is the operator's
 768 | # responsibility.
 769 | incremental_backups: false
 770 | 
 771 | # Whether or not to take a snapshot before each compaction.  Be
 772 | # careful using this option, since Cassandra won't clean up the
 773 | # snapshots for you.  Mostly useful if you're paranoid when there
 774 | # is a data format change.
 775 | snapshot_before_compaction: false
 776 | 
 777 | # Whether or not a snapshot is taken of the data before keyspace truncation
 778 | # or dropping of column families. The STRONGLY advised default of true 
 779 | # should be used to provide data safety. If you set this flag to false, you will
 780 | # lose data on truncation or drop.
 781 | auto_snapshot: true
 782 | 
 783 | # Granularity of the collation index of rows within a partition.
 784 | # Increase if your rows are large, or if you have a very large
 785 | # number of rows per partition.  The competing goals are these:
 786 | #
 787 | # - a smaller granularity means more index entries are generated
 788 | #   and looking up rows withing the partition by collation column
 789 | #   is faster
 790 | # - but, Cassandra will keep the collation index in memory for hot
 791 | #   rows (as part of the key cache), so a larger granularity means
 792 | #   you can cache more hot rows
 793 | column_index_size_in_kb: 64
 794 | 
 795 | # Per sstable indexed key cache entries (the collation index in memory
 796 | # mentioned above) exceeding this size will not be held on heap.
 797 | # This means that only partition information is held on heap and the
 798 | # index entries are read from disk.
 799 | #
 800 | # Note that this size refers to the size of the
 801 | # serialized index information and not the size of the partition.
 802 | column_index_cache_size_in_kb: 2
 803 | 
 804 | # Number of simultaneous compactions to allow, NOT including
 805 | # validation "compactions" for anti-entropy repair.  Simultaneous
 806 | # compactions can help preserve read performance in a mixed read/write
 807 | # workload, by mitigating the tendency of small sstables to accumulate
 808 | # during a single long running compactions. The default is usually
 809 | # fine and if you experience problems with compaction running too
 810 | # slowly or too fast, you should look at
 811 | # compaction_throughput_mb_per_sec first.
 812 | #
 813 | # concurrent_compactors defaults to the smaller of (number of disks,
 814 | # number of cores), with a minimum of 2 and a maximum of 8.
 815 | # 
 816 | # If your data directories are backed by SSD, you should increase this
 817 | # to the number of cores.
 818 | #concurrent_compactors: 1
 819 | 
 820 | # Throttles compaction to the given total throughput across the entire
 821 | # system. The faster you insert data, the faster you need to compact in
 822 | # order to keep the sstable count down, but in general, setting this to
 823 | # 16 to 32 times the rate you are inserting data is more than sufficient.
 824 | # Setting this to 0 disables throttling. Note that this account for all types
 825 | # of compaction, including validation compaction.
 826 | compaction_throughput_mb_per_sec: 16
 827 | 
 828 | # When compacting, the replacement sstable(s) can be opened before they
 829 | # are completely written, and used in place of the prior sstables for
 830 | # any range that has been written. This helps to smoothly transfer reads 
 831 | # between the sstables, reducing page cache churn and keeping hot rows hot
 832 | sstable_preemptive_open_interval_in_mb: 50
 833 | 
 834 | # Throttles all outbound streaming file transfers on this node to the
 835 | # given total throughput in Mbps. This is necessary because Cassandra does
 836 | # mostly sequential IO when streaming data during bootstrap or repair, which
 837 | # can lead to saturating the network connection and degrading rpc performance.
 838 | # When unset, the default is 200 Mbps or 25 MB/s.
 839 | # stream_throughput_outbound_megabits_per_sec: 200
 840 | 
 841 | # Throttles all streaming file transfer between the datacenters,
 842 | # this setting allows users to throttle inter dc stream throughput in addition
 843 | # to throttling all network stream traffic as configured with
 844 | # stream_throughput_outbound_megabits_per_sec
 845 | # When unset, the default is 200 Mbps or 25 MB/s
 846 | # inter_dc_stream_throughput_outbound_megabits_per_sec: 200
 847 | 
 848 | # How long the coordinator should wait for read operations to complete
 849 | read_request_timeout_in_ms: 5000
 850 | # How long the coordinator should wait for seq or index scans to complete
 851 | range_request_timeout_in_ms: 10000
 852 | # How long the coordinator should wait for writes to complete
 853 | write_request_timeout_in_ms: 2000
 854 | # How long the coordinator should wait for counter writes to complete
 855 | counter_write_request_timeout_in_ms: 5000
 856 | # How long a coordinator should continue to retry a CAS operation
 857 | # that contends with other proposals for the same row
 858 | cas_contention_timeout_in_ms: 1000
 859 | # How long the coordinator should wait for truncates to complete
 860 | # (This can be much longer, because unless auto_snapshot is disabled
 861 | # we need to flush first so we can snapshot before removing the data.)
 862 | truncate_request_timeout_in_ms: 60000
 863 | # The default timeout for other, miscellaneous operations
 864 | request_timeout_in_ms: 10000
 865 | 
 866 | # How long before a node logs slow queries. Select queries that take longer than
 867 | # this timeout to execute, will generate an aggregated log message, so that slow queries
 868 | # can be identified. Set this value to zero to disable slow query logging.
 869 | slow_query_log_timeout_in_ms: 500
 870 | 
 871 | # Enable operation timeout information exchange between nodes to accurately
 872 | # measure request timeouts.  If disabled, replicas will assume that requests
 873 | # were forwarded to them instantly by the coordinator, which means that
 874 | # under overload conditions we will waste that much extra time processing 
 875 | # already-timed-out requests.
 876 | #
 877 | # Warning: before enabling this property make sure to ntp is installed
 878 | # and the times are synchronized between the nodes.
 879 | cross_node_timeout: false
 880 | 
 881 | # Set keep-alive period for streaming
 882 | # This node will send a keep-alive message periodically with this period.
 883 | # If the node does not receive a keep-alive message from the peer for
 884 | # 2 keep-alive cycles the stream session times out and fail
 885 | # Default value is 300s (5 minutes), which means stalled stream
 886 | # times out in 10 minutes by default
 887 | # streaming_keep_alive_period_in_secs: 300
 888 | 
 889 | # phi value that must be reached for a host to be marked down.
 890 | # most users should never need to adjust this.
 891 | # phi_convict_threshold: 8
 892 | 
 893 | # endpoint_snitch -- Set this to a class that implements
 894 | # IEndpointSnitch.  The snitch has two functions:
 895 | #
 896 | # - it teaches Cassandra enough about your network topology to route
 897 | #   requests efficiently
 898 | # - it allows Cassandra to spread replicas around your cluster to avoid
 899 | #   correlated failures. It does this by grouping machines into
 900 | #   "datacenters" and "racks."  Cassandra will do its best not to have
 901 | #   more than one replica on the same "rack" (which may not actually
 902 | #   be a physical location)
 903 | #
 904 | # CASSANDRA WILL NOT ALLOW YOU TO SWITCH TO AN INCOMPATIBLE SNITCH
 905 | # ONCE DATA IS INSERTED INTO THE CLUSTER.  This would cause data loss.
 906 | # This means that if you start with the default SimpleSnitch, which
 907 | # locates every node on "rack1" in "datacenter1", your only options
 908 | # if you need to add another datacenter are GossipingPropertyFileSnitch
 909 | # (and the older PFS).  From there, if you want to migrate to an
 910 | # incompatible snitch like Ec2Snitch you can do it by adding new nodes
 911 | # under Ec2Snitch (which will locate them in a new "datacenter") and
 912 | # decommissioning the old ones.
 913 | #
 914 | # Out of the box, Cassandra provides:
 915 | #
 916 | # SimpleSnitch:
 917 | #    Treats Strategy order as proximity. This can improve cache
 918 | #    locality when disabling read repair.  Only appropriate for
 919 | #    single-datacenter deployments.
 920 | #
 921 | # GossipingPropertyFileSnitch
 922 | #    This should be your go-to snitch for production use.  The rack
 923 | #    and datacenter for the local node are defined in
 924 | #    cassandra-rackdc.properties and propagated to other nodes via
 925 | #    gossip.  If cassandra-topology.properties exists, it is used as a
 926 | #    fallback, allowing migration from the PropertyFileSnitch.
 927 | #
 928 | # PropertyFileSnitch:
 929 | #    Proximity is determined by rack and data center, which are
 930 | #    explicitly configured in cassandra-topology.properties.
 931 | #
 932 | # Ec2Snitch:
 933 | #    Appropriate for EC2 deployments in a single Region. Loads Region
 934 | #    and Availability Zone information from the EC2 API. The Region is
 935 | #    treated as the datacenter, and the Availability Zone as the rack.
 936 | #    Only private IPs are used, so this will not work across multiple
 937 | #    Regions.
 938 | #
 939 | # Ec2MultiRegionSnitch:
 940 | #    Uses public IPs as broadcast_address to allow cross-region
 941 | #    connectivity.  (Thus, you should set seed addresses to the public
 942 | #    IP as well.) You will need to open the storage_port or
 943 | #    ssl_storage_port on the public IP firewall.  (For intra-Region
 944 | #    traffic, Cassandra will switch to the private IP after
 945 | #    establishing a connection.)
 946 | #
 947 | # RackInferringSnitch:
 948 | #    Proximity is determined by rack and data center, which are
 949 | #    assumed to correspond to the 3rd and 2nd octet of each node's IP
 950 | #    address, respectively.  Unless this happens to match your
 951 | #    deployment conventions, this is best used as an example of
 952 | #    writing a custom Snitch class and is provided in that spirit.
 953 | #
 954 | # You can use a custom Snitch by setting this to the full class name
 955 | # of the snitch, which will be assumed to be on your classpath.
 956 | endpoint_snitch: SimpleSnitch
 957 | 
 958 | # controls how often to perform the more expensive part of host score
 959 | # calculation
 960 | dynamic_snitch_update_interval_in_ms: 100 
 961 | # controls how often to reset all host scores, allowing a bad host to
 962 | # possibly recover
 963 | dynamic_snitch_reset_interval_in_ms: 600000
 964 | # if set greater than zero and read_repair_chance is < 1.0, this will allow
 965 | # 'pinning' of replicas to hosts in order to increase cache capacity.
 966 | # The badness threshold will control how much worse the pinned host has to be
 967 | # before the dynamic snitch will prefer other replicas over it.  This is
 968 | # expressed as a double which represents a percentage.  Thus, a value of
 969 | # 0.2 means Cassandra would continue to prefer the static snitch values
 970 | # until the pinned host was 20% worse than the fastest.
 971 | dynamic_snitch_badness_threshold: 0.1
 972 | 
 973 | # request_scheduler -- Set this to a class that implements
 974 | # RequestScheduler, which will schedule incoming client requests
 975 | # according to the specific policy. This is useful for multi-tenancy
 976 | # with a single Cassandra cluster.
 977 | # NOTE: This is specifically for requests from the client and does
 978 | # not affect inter node communication.
 979 | # org.apache.cassandra.scheduler.NoScheduler - No scheduling takes place
 980 | # org.apache.cassandra.scheduler.RoundRobinScheduler - Round robin of
 981 | # client requests to a node with a separate queue for each
 982 | # request_scheduler_id. The scheduler is further customized by
 983 | # request_scheduler_options as described below.
 984 | request_scheduler: org.apache.cassandra.scheduler.NoScheduler
 985 | 
 986 | # Scheduler Options vary based on the type of scheduler
 987 | #
 988 | # NoScheduler
 989 | #   Has no options
 990 | #
 991 | # RoundRobin
 992 | #   throttle_limit
 993 | #     The throttle_limit is the number of in-flight
 994 | #     requests per client.  Requests beyond 
 995 | #     that limit are queued up until
 996 | #     running requests can complete.
 997 | #     The value of 80 here is twice the number of
 998 | #     concurrent_reads + concurrent_writes.
 999 | #   default_weight
1000 | #     default_weight is optional and allows for
1001 | #     overriding the default which is 1.
1002 | #   weights
1003 | #     Weights are optional and will default to 1 or the
1004 | #     overridden default_weight. The weight translates into how
1005 | #     many requests are handled during each turn of the
1006 | #     RoundRobin, based on the scheduler id.
1007 | #
1008 | # request_scheduler_options:
1009 | #    throttle_limit: 80
1010 | #    default_weight: 5
1011 | #    weights:
1012 | #      Keyspace1: 1
1013 | #      Keyspace2: 5
1014 | 
1015 | # request_scheduler_id -- An identifier based on which to perform
1016 | # the request scheduling. Currently the only valid option is keyspace.
1017 | # request_scheduler_id: keyspace
1018 | 
1019 | # Enable or disable inter-node encryption
1020 | # JVM defaults for supported SSL socket protocols and cipher suites can
1021 | # be replaced using custom encryption options. This is not recommended
1022 | # unless you have policies in place that dictate certain settings, or
1023 | # need to disable vulnerable ciphers or protocols in case the JVM cannot
1024 | # be updated.
1025 | # FIPS compliant settings can be configured at JVM level and should not
1026 | # involve changing encryption settings here:
1027 | # https://docs.oracle.com/javase/8/docs/technotes/guides/security/jsse/FIPS.html
1028 | # *NOTE* No custom encryption options are enabled at the moment
1029 | # The available internode options are : all, none, dc, rack
1030 | #
1031 | # If set to dc cassandra will encrypt the traffic between the DCs
1032 | # If set to rack cassandra will encrypt the traffic between the racks
1033 | #
1034 | # The passwords used in these options must match the passwords used when generating
1035 | # the keystore and truststore.  For instructions on generating these files, see:
1036 | # http://download.oracle.com/javase/6/docs/technotes/guides/security/jsse/JSSERefGuide.html#CreateKeystore
1037 | #
1038 | server_encryption_options:
1039 |     internode_encryption: none
1040 |     keystore: conf/.keystore
1041 |     keystore_password: cassandra
1042 |     truststore: conf/.truststore
1043 |     truststore_password: cassandra
1044 |     # More advanced defaults below:
1045 |     # protocol: TLS
1046 |     # algorithm: SunX509
1047 |     # store_type: JKS
1048 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1049 |     # require_client_auth: false
1050 |     # require_endpoint_verification: false
1051 | 
1052 | # enable or disable client/server encryption.
1053 | client_encryption_options:
1054 |     enabled: false
1055 |     # If enabled and optional is set to true encrypted and unencrypted connections are handled.
1056 |     optional: false
1057 |     keystore: conf/.keystore
1058 |     keystore_password: cassandra
1059 |     # require_client_auth: false
1060 |     # Set trustore and truststore_password if require_client_auth is true
1061 |     # truststore: conf/.truststore
1062 |     # truststore_password: cassandra
1063 |     # More advanced defaults below:
1064 |     # protocol: TLS
1065 |     # algorithm: SunX509
1066 |     # store_type: JKS
1067 |     # cipher_suites: [TLS_RSA_WITH_AES_128_CBC_SHA,TLS_RSA_WITH_AES_256_CBC_SHA,TLS_DHE_RSA_WITH_AES_128_CBC_SHA,TLS_DHE_RSA_WITH_AES_256_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA]
1068 | 
1069 | # internode_compression controls whether traffic between nodes is
1070 | # compressed.
1071 | # Can be:
1072 | #
1073 | # all
1074 | #   all traffic is compressed
1075 | #
1076 | # dc
1077 | #   traffic between different datacenters is compressed
1078 | #
1079 | # none
1080 | #   nothing is compressed.
1081 | internode_compression: dc
1082 | 
1083 | # Enable or disable tcp_nodelay for inter-dc communication.
1084 | # Disabling it will result in larger (but fewer) network packets being sent,
1085 | # reducing overhead from the TCP protocol itself, at the cost of increasing
1086 | # latency if you block for cross-datacenter responses.
1087 | inter_dc_tcp_nodelay: false
1088 | 
1089 | # TTL for different trace types used during logging of the repair process.
1090 | tracetype_query_ttl: 86400
1091 | tracetype_repair_ttl: 604800
1092 | 
1093 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1094 | # This threshold can be adjusted to minimize logging if necessary
1095 | # gc_log_threshold_in_ms: 200
1096 | 
1097 | # If unset, all GC Pauses greater than gc_log_threshold_in_ms will log at
1098 | # INFO level
1099 | # UDFs (user defined functions) are disabled by default.
1100 | # As of Cassandra 3.0 there is a sandbox in place that should prevent execution of evil code.
1101 | enable_user_defined_functions: false
1102 | 
1103 | # Enables scripted UDFs (JavaScript UDFs).
1104 | # Java UDFs are always enabled, if enable_user_defined_functions is true.
1105 | # Enable this option to be able to use UDFs with "language javascript" or any custom JSR-223 provider.
1106 | # This option has no effect, if enable_user_defined_functions is false.
1107 | enable_scripted_user_defined_functions: false
1108 | 
1109 | # The default Windows kernel timer and scheduling resolution is 15.6ms for power conservation.
1110 | # Lowering this value on Windows can provide much tighter latency and better throughput, however
1111 | # some virtualized environments may see a negative performance impact from changing this setting
1112 | # below their system default. The sysinternals 'clockres' tool can confirm your system's default
1113 | # setting.
1114 | windows_timer_interval: 1
1115 | 
1116 | 
1117 | # Enables encrypting data at-rest (on disk). Different key providers can be plugged in, but the default reads from
1118 | # a JCE-style keystore. A single keystore can hold multiple keys, but the one referenced by
1119 | # the "key_alias" is the only key that will be used for encrypt opertaions; previously used keys
1120 | # can still (and should!) be in the keystore and will be used on decrypt operations
1121 | # (to handle the case of key rotation).
1122 | #
1123 | # It is strongly recommended to download and install Java Cryptography Extension (JCE)
1124 | # Unlimited Strength Jurisdiction Policy Files for your version of the JDK.
1125 | # (current link: http://www.oracle.com/technetwork/java/javase/downloads/jce8-download-2133166.html)
1126 | #
1127 | # Currently, only the following file types are supported for transparent data encryption, although
1128 | # more are coming in future cassandra releases: commitlog, hints
1129 | transparent_data_encryption_options:
1130 |     enabled: false
1131 |     chunk_length_kb: 64
1132 |     cipher: AES/CBC/PKCS5Padding
1133 |     key_alias: testing:1
1134 |     # CBC IV length for AES needs to be 16 bytes (which is also the default size)
1135 |     # iv_length: 16
1136 |     key_provider: 
1137 |       - class_name: org.apache.cassandra.security.JKSKeyProvider
1138 |         parameters: 
1139 |           - keystore: conf/.keystore
1140 |             keystore_password: cassandra
1141 |             store_type: JCEKS
1142 |             key_password: cassandra
1143 | 
1144 | 
1145 | #####################
1146 | # SAFETY THRESHOLDS #
1147 | #####################
1148 | 
1149 | # When executing a scan, within or across a partition, we need to keep the
1150 | # tombstones seen in memory so we can return them to the coordinator, which
1151 | # will use them to make sure other replicas also know about the deleted rows.
1152 | # With workloads that generate a lot of tombstones, this can cause performance
1153 | # problems and even exaust the server heap.
1154 | # (http://www.datastax.com/dev/blog/cassandra-anti-patterns-queues-and-queue-like-datasets)
1155 | # Adjust the thresholds here if you understand the dangers and want to
1156 | # scan more tombstones anyway.  These thresholds may also be adjusted at runtime
1157 | # using the StorageService mbean.
1158 | tombstone_warn_threshold: 1000
1159 | tombstone_failure_threshold: 100000
1160 | 
1161 | # Log WARN on any multiple-partition batch size exceeding this value. 5kb per batch by default.
1162 | # Caution should be taken on increasing the size of this threshold as it can lead to node instability.
1163 | batch_size_warn_threshold_in_kb: 5
1164 | 
1165 | # Fail any multiple-partition batch exceeding this value. 50kb (10x warn threshold) by default.
1166 | batch_size_fail_threshold_in_kb: 50
1167 | 
1168 | # Log WARN on any batches not of type LOGGED than span across more partitions than this limit
1169 | unlogged_batch_across_partitions_warn_threshold: 10
1170 | 
1171 | # Log a warning when compacting partitions larger than this value
1172 | compaction_large_partition_warning_threshold_mb: 100
1173 | 
1174 | # GC Pauses greater than gc_warn_threshold_in_ms will be logged at WARN level
1175 | # Adjust the threshold based on your application throughput requirement
1176 | # By default, Cassandra logs GC Pauses greater than 200 ms at INFO level
1177 | gc_warn_threshold_in_ms: 1000
1178 | 
1179 | # Maximum size of any value in SSTables. Safety measure to detect SSTable corruption
1180 | # early. Any value size larger than this threshold will result into marking an SSTable
1181 | # as corrupted.
1182 | # max_value_size_in_mb: 256
1183 | 
1184 | # Back-pressure settings #
1185 | # If enabled, the coordinator will apply the back-pressure strategy specified below to each mutation
1186 | # sent to replicas, with the aim of reducing pressure on overloaded replicas.
1187 | back_pressure_enabled: false
1188 | # The back-pressure strategy applied.
1189 | # The default implementation, RateBasedBackPressure, takes three arguments:
1190 | # high ratio, factor, and flow type, and uses the ratio between incoming mutation responses and outgoing mutation requests.
1191 | # If below high ratio, outgoing mutations are rate limited according to the incoming rate decreased by the given factor;
1192 | # if above high ratio, the rate limiting is increased by the given factor;
1193 | # such factor is usually best configured between 1 and 10, use larger values for a faster recovery
1194 | # at the expense of potentially more dropped mutations;
1195 | # the rate limiting is applied according to the flow type: if FAST, it's rate limited at the speed of the fastest replica,
1196 | # if SLOW at the speed of the slowest one.
1197 | # New strategies can be added. Implementors need to implement org.apache.cassandra.net.BackpressureStrategy and
1198 | # provide a public constructor accepting a Map<String, Object>.
1199 | back_pressure_strategy:
1200 |     - class_name: org.apache.cassandra.net.RateBasedBackPressure
1201 |       parameters:
1202 |         - high_ratio: 0.90
1203 |           factor: 5
1204 |           flow: FAST
1205 | 
1206 | # Coalescing Strategies #
1207 | # Coalescing multiples messages turns out to significantly boost message processing throughput (think doubling or more).
1208 | # On bare metal, the floor for packet processing throughput is high enough that many applications won't notice, but in
1209 | # virtualized environments, the point at which an application can be bound by network packet processing can be
1210 | # surprisingly low compared to the throughput of task processing that is possible inside a VM. It's not that bare metal
1211 | # doesn't benefit from coalescing messages, it's that the number of packets a bare metal network interface can process
1212 | # is sufficient for many applications such that no load starvation is experienced even without coalescing.
1213 | # There are other benefits to coalescing network messages that are harder to isolate with a simple metric like messages
1214 | # per second. By coalescing multiple tasks together, a network thread can process multiple messages for the cost of one
1215 | # trip to read from a socket, and all the task submission work can be done at the same time reducing context switching
1216 | # and increasing cache friendliness of network message processing.
1217 | # See CASSANDRA-8692 for details.
1218 | 
1219 | # Strategy to use for coalescing messages in OutboundTcpConnection.
1220 | # Can be fixed, movingaverage, timehorizon, disabled (default).
1221 | # You can also specify a subclass of CoalescingStrategies.CoalescingStrategy by name.
1222 | # otc_coalescing_strategy: DISABLED
1223 | 
1224 | # How many microseconds to wait for coalescing. For fixed strategy this is the amount of time after the first
1225 | # message is received before it will be sent with any accompanying messages. For moving average this is the
1226 | # maximum amount of time that will be waited as well as the interval at which messages must arrive on average
1227 | # for coalescing to be enabled.
1228 | # otc_coalescing_window_us: 200
1229 | 
1230 | # Do not try to coalesce messages if we already got that many messages. This should be more than 2 and less than 128.
1231 | # otc_coalescing_enough_coalesced_messages: 8
1232 | 
1233 | # How many milliseconds to wait between two expiration runs on the backlog (queue) of the OutboundTcpConnection.
1234 | # Expiration is done if messages are piling up in the backlog. Droppable messages are expired to free the memory
1235 | # taken by expired messages. The interval should be between 0 and 1000, and in most installations the default value
1236 | # will be appropriate. A smaller value could potentially expire messages slightly sooner at the expense of more CPU
1237 | # time and queue contention while iterating the backlog of messages.
1238 | # An interval of 0 disables any wait time, which is the behavior of former Cassandra versions.
1239 | #
1240 | # otc_backlog_expiration_interval_ms: 200
1241 | 


--------------------------------------------------------------------------------
/cassandra-cdc/pom.xml:
--------------------------------------------------------------------------------
  1 | <?xml version="1.0" encoding="UTF-8"?>
  2 | <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  3 |     xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
  4 |     <modelVersion>4.0.0</modelVersion>
  5 | 
  6 |     <groupId>io.smartcat</groupId>
  7 |     <artifactId>cassandra-cdc</artifactId>
  8 |     <version>0.0.1-SNAPSHOT</version>
  9 |     <packaging>jar</packaging>
 10 | 
 11 |     <name>CassandraCDC</name>
 12 |     <description>Custom Cassandra CDC Reader</description>
 13 | 
 14 | 
 15 |     <properties>
 16 |         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
 17 |         <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
 18 |         <source.level>1.8</source.level>
 19 |         <code.level>1.8</code.level>
 20 |         <version.cassandra-all>3.11.0</version.cassandra-all>
 21 |         <version.kafka>0.10.2.0</version.kafka>
 22 |         <version.plugin.compiler>3.5.1</version.plugin.compiler>
 23 |         <version.plugin.install>2.5.2</version.plugin.install>
 24 |         <version.plugin.deploy>2.8.2</version.plugin.deploy>
 25 |         <version.plugin.shade>2.4.3</version.plugin.shade>
 26 |         <version.plugin.jar>2.6</version.plugin.jar>
 27 |         <version.plugin.resources>2.7</version.plugin.resources>
 28 |     </properties>
 29 | 
 30 |     <dependencies>
 31 |         <dependency>
 32 |             <groupId>org.apache.cassandra</groupId>
 33 |             <artifactId>cassandra-all</artifactId>
 34 |             <version>${version.cassandra-all}</version>
 35 |         </dependency>
 36 |         <dependency>
 37 |             <groupId>org.apache.kafka</groupId>
 38 |             <artifactId>connect-api</artifactId>
 39 |             <version>${version.kafka}</version>
 40 |         </dependency>
 41 |         <dependency>
 42 |             <groupId>org.apache.kafka</groupId>
 43 |             <artifactId>connect-runtime</artifactId>
 44 |             <version>${version.kafka}</version>
 45 |             <exclusions>
 46 |                 <exclusion>
 47 |                     <groupId>org.slf4j</groupId>
 48 |                     <artifactId>slf4j-log4j12</artifactId>
 49 |                 </exclusion>
 50 |             </exclusions>
 51 |         </dependency>
 52 |     </dependencies>
 53 | 
 54 |     <build>
 55 |         <pluginManagement>
 56 |             <plugins>
 57 |                 <plugin>
 58 |                     <groupId>org.apache.maven.plugins</groupId>
 59 |                     <artifactId>maven-jar-plugin</artifactId>
 60 |                     <version>${version.plugin.jar}</version>
 61 |                 </plugin>
 62 |                 <plugin>
 63 |                     <groupId>org.apache.maven.plugins</groupId>
 64 |                     <artifactId>maven-resources-plugin</artifactId>
 65 |                     <version>${version.plugin.resources}</version>
 66 |                 </plugin>
 67 | 
 68 |                 <plugin>
 69 |                     <groupId>org.apache.maven.plugins</groupId>
 70 |                     <artifactId>maven-install-plugin</artifactId>
 71 |                     <version>${version.plugin.install}</version>
 72 |                 </plugin>
 73 |                 <plugin>
 74 |                     <groupId>org.apache.maven.plugins</groupId>
 75 |                     <artifactId>maven-shade-plugin</artifactId>
 76 |                     <version>${version.plugin.shade}</version>
 77 |                 </plugin>
 78 |                 <plugin>
 79 |                     <groupId>org.apache.maven.plugins</groupId>
 80 |                     <artifactId>maven-deploy-plugin</artifactId>
 81 |                     <version>${version.plugin.deploy}</version>
 82 |                     <configuration>
 83 |                         <updateReleaseInfo>true</updateReleaseInfo>
 84 |                     </configuration>
 85 |                 </plugin>
 86 |                 <plugin>
 87 |                     <groupId>org.apache.maven.plugins</groupId>
 88 |                     <artifactId>maven-compiler-plugin</artifactId>
 89 |                     <version>${version.plugin.compiler}</version>
 90 |                     <inherited>true</inherited>
 91 |                     <configuration>
 92 |                         <source>${source.level}</source>
 93 |                         <target>${code.level}</target>
 94 |                         <encoding>${project.build.sourceEncoding}</encoding>
 95 |                     </configuration>
 96 |                 </plugin>
 97 |             </plugins>
 98 |         </pluginManagement>
 99 |         <plugins>
100 |             <plugin>
101 |                 <groupId>org.apache.maven.plugins</groupId>
102 |                 <artifactId>maven-shade-plugin</artifactId>
103 |                 <executions>
104 |                     <execution>
105 |                         <phase>package</phase>
106 |                         <goals>
107 |                             <goal>shade</goal>
108 |                         </goals>
109 |                         <configuration>
110 |                             <shadedArtifactAttached>false</shadedArtifactAttached>
111 |                             <createDependencyReducedPom>false</createDependencyReducedPom>
112 |                             <artifactSet>
113 |                                 <excludes>
114 |                                   <exclude>classworlds:classworlds</exclude>
115 |                                   <exclude>junit:junit</exclude>
116 |                                   <exclude>org.apache.maven:lib:tests</exclude>
117 |                                 </excludes>
118 |                             </artifactSet>
119 |                             <transformers>
120 |                                 <transformer
121 |                                     implementation="org.apache.maven.plugins.shade.resource.ManifestResourceTransformer">
122 |                                     <mainClass>io.smartcat.cassandra.cdc.Reader</mainClass>
123 |                                 </transformer>
124 |                             </transformers>
125 |                         </configuration>
126 |                     </execution>
127 |                 </executions>
128 |             </plugin>
129 |         </plugins>
130 |     </build>
131 | </project>
132 | 


--------------------------------------------------------------------------------
/cassandra-cdc/src/main/java/io/smartcat/cassandra/cdc/CustomCommitLogReadHandler.java:
--------------------------------------------------------------------------------
  1 | package io.smartcat.cassandra.cdc;
  2 | 
  3 | import org.apache.cassandra.config.ColumnDefinition;
  4 | import org.apache.cassandra.db.Clustering;
  5 | import org.apache.cassandra.db.ClusteringBound;
  6 | import org.apache.cassandra.db.ClusteringPrefix;
  7 | import org.apache.cassandra.db.Mutation;
  8 | import org.apache.cassandra.db.commitlog.CommitLogDescriptor;
  9 | import org.apache.cassandra.db.commitlog.CommitLogReadHandler;
 10 | import org.apache.cassandra.db.partitions.Partition;
 11 | import org.apache.cassandra.db.partitions.PartitionUpdate;
 12 | import org.apache.cassandra.db.rows.Cell;
 13 | import org.apache.cassandra.db.rows.Row;
 14 | import org.apache.cassandra.db.rows.Unfiltered;
 15 | import org.apache.cassandra.db.rows.UnfilteredRowIterator;
 16 | import org.apache.kafka.clients.producer.KafkaProducer;
 17 | import org.apache.kafka.clients.producer.Producer;
 18 | import org.apache.kafka.clients.producer.ProducerRecord;
 19 | import org.json.simple.JSONObject;
 20 | import org.slf4j.Logger;
 21 | import org.slf4j.LoggerFactory;
 22 | 
 23 | import java.io.IOException;
 24 | import java.util.ArrayList;
 25 | import java.util.Iterator;
 26 | import java.util.List;
 27 | import java.util.Map;
 28 | 
 29 | public class CustomCommitLogReadHandler implements CommitLogReadHandler {
 30 | 
 31 |     private static final Logger LOGGER = LoggerFactory.getLogger(CustomCommitLogReadHandler.class);
 32 | 
 33 |     private final String keyspace;
 34 |     private final String table;
 35 |     private final String topic;
 36 |     private final Producer<String, String> producer;
 37 | 
 38 |     public CustomCommitLogReadHandler(Map<String, Object> configuration) {
 39 |         keyspace = (String) YamlUtils.select(configuration, "cassandra.keyspace");
 40 |         table = (String) YamlUtils.select(configuration, "cassandra.table");
 41 |         topic = (String) YamlUtils.select(configuration, "kafka.topic");
 42 |         producer = new KafkaProducer<>((Map) YamlUtils.select(configuration, "kafka.configuration"));
 43 |     }
 44 | 
 45 |     @Override
 46 |     public void handleMutation(Mutation mutation, int size, int entryLocation, CommitLogDescriptor descriptor) {
 47 |         LOGGER.debug("Handle mutation started...");
 48 |         for (PartitionUpdate partitionUpdate : mutation.getPartitionUpdates()) {
 49 |             process(partitionUpdate);
 50 |         }
 51 |         LOGGER.debug("Handle mutation finished...");
 52 |     }
 53 | 
 54 |     @Override
 55 |     public void handleUnrecoverableError(CommitLogReadException exception) throws IOException {
 56 |         LOGGER.debug("Handle unrecoverable error called.");
 57 |         throw new RuntimeException(exception);
 58 |     }
 59 | 
 60 |     @Override
 61 |     public boolean shouldSkipSegmentOnError(CommitLogReadException exception) throws IOException {
 62 |         LOGGER.debug("Should skip segment on error.");
 63 |         exception.printStackTrace();
 64 |         return true;
 65 |     }
 66 | 
 67 | 
 68 |     @SuppressWarnings("unchecked")
 69 |     private void process(Partition partition) {
 70 |         LOGGER.debug("Process method started...");
 71 |         if (!partition.metadata().ksName.equals(keyspace)) {
 72 |             LOGGER.debug("Keyspace should be '{}' but is '{}'.", keyspace, partition.metadata().ksName);
 73 |             return;
 74 |         }
 75 |         if (!partition.metadata().cfName.equals(table)) {
 76 |             LOGGER.debug("Table should be '{} but is '{}'.", table, partition.metadata().cfName);
 77 |             return;
 78 |         }
 79 |         String key = getKey(partition);
 80 |         JSONObject obj = new JSONObject();
 81 |         obj.put("key", key);
 82 |         if (partitionIsDeleted(partition)) {
 83 |             obj.put("partitionDeleted", true);
 84 |         } else {
 85 |             UnfilteredRowIterator it = partition.unfilteredIterator();
 86 |             List<JSONObject> rows = new ArrayList<>();
 87 |             while (it.hasNext()) {
 88 |                 Unfiltered un = it.next();
 89 |                 if (un.isRow()) {
 90 |                     JSONObject jsonRow = new JSONObject();
 91 |                     Clustering clustering = (Clustering) un.clustering();
 92 |                     String clusteringKey = clustering.toCQLString(partition.metadata());
 93 |                     jsonRow.put("clusteringKey", clusteringKey);
 94 |                     Row row = partition.getRow(clustering);
 95 | 
 96 |                     if (rowIsDeleted(row)) {
 97 |                         obj.put("rowDeleted", true);
 98 |                     } else {
 99 |                         Iterator<Cell> cells = row.cells().iterator();
100 |                         Iterator<ColumnDefinition> columns = row.columns().iterator();
101 |                         List<JSONObject> cellObjects = new ArrayList<>();
102 |                         while (cells.hasNext() && columns.hasNext()) {
103 |                             JSONObject jsonCell = new JSONObject();
104 |                             ColumnDefinition columnDef = columns.next();
105 |                             Cell cell = cells.next();
106 |                             jsonCell.put("name", columnDef.name.toString());
107 |                             if (cell.isTombstone()) {
108 |                                 jsonCell.put("deleted", true);
109 |                             } else {
110 |                                 String data = columnDef.type.getString(cell.value());
111 |                                 jsonCell.put("value", data);
112 |                             }
113 |                             cellObjects.add(jsonCell);
114 |                         }
115 |                         jsonRow.put("cells", cellObjects);
116 |                     }
117 |                     rows.add(jsonRow);
118 |                 } else if (un.isRangeTombstoneMarker()) {
119 |                     obj.put("rowRangeDeleted", true);
120 |                     ClusteringBound bound = (ClusteringBound) un.clustering();
121 |                     List<JSONObject> bounds = new ArrayList<>();
122 |                     for (int i = 0; i < bound.size(); i++) {
123 |                         String clusteringBound = partition.metadata().comparator.subtype(i).getString(bound.get(i));
124 |                         JSONObject boundObject = new JSONObject();
125 |                         boundObject.put("clusteringKey", clusteringBound);
126 |                         if (i == bound.size() - 1) {
127 |                             if (bound.kind().isStart()) {
128 |                                 boundObject.put("inclusive",
129 |                                         bound.kind() == ClusteringPrefix.Kind.INCL_START_BOUND ? true : false);
130 |                             }
131 |                             if (bound.kind().isEnd()) {
132 |                                 boundObject.put("inclusive",
133 |                                         bound.kind() == ClusteringPrefix.Kind.INCL_END_BOUND ? true : false);
134 |                             }
135 |                         }
136 |                         bounds.add(boundObject);
137 |                     }
138 |                     obj.put((bound.kind().isStart() ? "start" : "end"), bounds);
139 |                 }
140 |             }
141 |             obj.put("rows", rows);
142 |         }
143 |         LOGGER.debug("Creating json value...");
144 |         String value = obj.toJSONString();
145 |         LOGGER.debug("Created json value '{}'", value);
146 |         ProducerRecord<String, String> record = new ProducerRecord<>(topic, key, value);
147 |         LOGGER.debug("Created producer record with topic {}, key {}, value {}", topic, key, value);
148 |         producer.send(record);
149 |         LOGGER.debug("Sent record to kafka.");
150 |     }
151 | 
152 |     private boolean partitionIsDeleted(Partition partition) {
153 |         return partition.partitionLevelDeletion().markedForDeleteAt() > Long.MIN_VALUE;
154 |     }
155 | 
156 |     private boolean rowIsDeleted(Row row) {
157 |         return row.deletion().time().markedForDeleteAt() > Long.MIN_VALUE;
158 |     }
159 | 
160 |     private String getKey(Partition partition) {
161 |         return partition.metadata().getKeyValidator().getString(partition.partitionKey().getKey());
162 |     }
163 | }
164 | 


--------------------------------------------------------------------------------
/cassandra-cdc/src/main/java/io/smartcat/cassandra/cdc/Reader.java:
--------------------------------------------------------------------------------
 1 | package io.smartcat.cassandra.cdc;
 2 | 
 3 | import org.apache.cassandra.config.DatabaseDescriptor;
 4 | import org.apache.cassandra.config.Schema;
 5 | import org.apache.cassandra.db.commitlog.CommitLogReader;
 6 | import org.slf4j.Logger;
 7 | import org.slf4j.LoggerFactory;
 8 | 
 9 | import java.io.IOException;
10 | import java.nio.file.*;
11 | import java.util.Map;
12 | 
13 | import static java.nio.file.StandardWatchEventKinds.ENTRY_CREATE;
14 | 
15 | public class Reader {
16 | 
17 |     private static final Logger LOGGER = LoggerFactory.getLogger(CustomCommitLogReadHandler.class);
18 | 
19 |     private final WatchService watcher;
20 |     private final Path dir;
21 |     private final WatchKey key;
22 |     private final CommitLogReader commitLogReader;
23 |     private final CustomCommitLogReadHandler commitLogReadHander;
24 | 
25 |     /**
26 |      * Creates a WatchService and registers the given directory
27 |      */
28 |     public Reader(Map<String, Object> configuration) throws IOException {
29 |         this.dir = Paths.get((String) YamlUtils.select(configuration, "cassandra.cdc_raw_directory"));
30 |         watcher = FileSystems.getDefault().newWatchService();
31 |         key = dir.register(watcher, ENTRY_CREATE);
32 |         commitLogReader = new CommitLogReader();
33 |         commitLogReadHander = new CustomCommitLogReadHandler(configuration);
34 |         DatabaseDescriptor.toolInitialization();
35 |         Schema.instance.loadFromDisk(false);
36 |     }
37 | 
38 |     /**
39 |      * Process all events for keys queued to the watcher
40 |      *
41 |      * @throws InterruptedException
42 |      * @throws IOException
43 |      */
44 |     public void processEvents() throws InterruptedException, IOException {
45 |         while (true) {
46 |             WatchKey aKey = watcher.take();
47 |             if (!key.equals(aKey)) {
48 |                 LOGGER.error("WatchKey not recognized.");
49 |                 continue;
50 |             }
51 |             for (WatchEvent<?> event : key.pollEvents()) {
52 |                 WatchEvent.Kind<?> kind = event.kind();
53 |                 if (kind != ENTRY_CREATE) {
54 |                     continue;
55 |                 }
56 | 
57 |                 // Context for directory entry event is the file name of entry
58 |                 WatchEvent<Path> ev = (WatchEvent<Path>) event;
59 |                 Path relativePath = ev.context();
60 |                 Path absolutePath = dir.resolve(relativePath);
61 |                 processCommitLogSegment(absolutePath);
62 |                 Files.delete(absolutePath);
63 | 
64 |                 // print out event
65 |                 LOGGER.debug("{}: {}", event.kind().name(), absolutePath);
66 |             }
67 |             key.reset();
68 |         }
69 |     }
70 | 
71 |     public static void main(String[] args) throws IOException, InterruptedException {
72 |         Map<String, Object> configuration = YamlUtils.load(args[0]);
73 |         new Reader(configuration).processEvents();
74 |     }
75 | 
76 |     private void processCommitLogSegment(Path path) throws IOException {
77 |         LOGGER.debug("Processing commitlog segment...");
78 |         commitLogReader.readCommitLogSegment(commitLogReadHander, path.toFile(), false);
79 |         LOGGER.debug("Commitlog segment processed.");
80 |     }
81 | }
82 | 


--------------------------------------------------------------------------------
/cassandra-cdc/src/main/java/io/smartcat/cassandra/cdc/YamlUtils.java:
--------------------------------------------------------------------------------
 1 | package io.smartcat.cassandra.cdc;
 2 | 
 3 | import org.apache.cassandra.io.util.FileUtils;
 4 | import org.yaml.snakeyaml.Yaml;
 5 | 
 6 | import java.io.File;
 7 | import java.io.FileInputStream;
 8 | import java.io.InputStream;
 9 | import java.util.ArrayList;
10 | import java.util.List;
11 | import java.util.Map;
12 | import java.util.regex.Matcher;
13 | import java.util.regex.Pattern;
14 | 
15 | public class YamlUtils {
16 | 
17 |     private static final Pattern PATH_INDEX = Pattern.compile("^(.+)\\[([0-9]+)\\]$");
18 | 
19 |     @SuppressWarnings("unchecked")
20 |     public static Map<String, Object> load(String filePath) {
21 |         InputStream stream = null;
22 |         try {
23 |             stream = new FileInputStream(new File(filePath));
24 |             Yaml yaml = new Yaml();
25 |             return (Map<String, Object>) yaml.load(stream);
26 |         } catch (Exception e) {
27 |             throw new RuntimeException(e);
28 |         } finally {
29 |             FileUtils.closeQuietly(stream);
30 |         }
31 |     }
32 | 
33 |     public static Object select(Map<String, Object> configuration, String path) {
34 |         List<Object> pathComponents = toPathList(path);
35 |         if (pathComponents.isEmpty()) {
36 |             throw new IllegalArgumentException(
37 |                     String.format("Given path expression %s has to have at least one component.", pathComponents));
38 |         }
39 |         return extractPathRecursive(configuration, pathComponents);
40 |     }
41 | 
42 |     private static Object extractPathRecursive(Object input, List<Object> path) {
43 |         if (path.isEmpty()) {
44 |             return input;
45 |         } else {
46 |             Object obj;
47 |             Object firstElement = path.get(0);
48 |             if (firstElement instanceof Integer) {
49 |                 Integer childIndex = (Integer) firstElement;
50 |                 obj = ((List<?>) input).get(childIndex);
51 |             } else {
52 |                 String childName = (String) firstElement;
53 |                 obj = ((Map<?, ?>) input).get(childName);
54 |             }
55 |             return extractPathRecursive(obj, path.subList(1, path.size()));
56 |         }
57 |     }
58 | 
59 |     private static List<Object> toPathList(String jsonPath) {
60 |         String[] pathComponents = jsonPath.split("\\.");
61 |         List<Object> pathList = new ArrayList<>();
62 |         for (String component : pathComponents) {
63 |             Matcher m = PATH_INDEX.matcher(component);
64 |             if (m.matches()) {
65 |                 pathList.add(m.group(1));
66 |                 pathList.add(Integer.valueOf(m.group(2)));
67 |             } else {
68 |                 pathList.add(component);
69 |             }
70 |         }
71 |         return pathList;
72 |     }
73 | }
74 | 


--------------------------------------------------------------------------------
/cassandra-cdc/src/main/resources/logback.xml:
--------------------------------------------------------------------------------
 1 | <?xml version="1.0" encoding="UTF-8"?>
 2 | <configuration>
 3 |     <property name="LOG_DIR" value="logs" />
 4 | 
 5 |     <appender name="STDOUT" class="ch.qos.logback.core.ConsoleAppender">
 6 |         <encoder>
 7 |             <pattern>%d{yyyy-MM-dd HH:mm:ss.SSS} %level [%thread] %logger{1} - %msg%n</pattern>
 8 |         </encoder>
 9 |     </appender>
10 | 
11 |     <appender name="ROLLING_FILE" class="ch.qos.logback.core.rolling.RollingFileAppender">
12 |         <file>${LOG_DIR}/berserker.log</file>
13 |         <encoder class="ch.qos.logback.classic.encoder.PatternLayoutEncoder">
14 |             <Pattern>%d{yyyy-MM-dd HH:mm:ss.SSS} %level [%thread] %logger{1} - %msg%n</Pattern>
15 |         </encoder>
16 |         <rollingPolicy class="ch.qos.logback.core.rolling.TimeBasedRollingPolicy">
17 |             <!-- rollover daily -->
18 |             <fileNamePattern>${LOG_DIR}/archived/berserker.%d{yyyy-MM-dd}.%i.log</fileNamePattern>
19 |             <timeBasedFileNamingAndTriggeringPolicy class="ch.qos.logback.core.rolling.SizeAndTimeBasedFNATP">
20 |                 <maxFileSize>10MB</maxFileSize>
21 |             </timeBasedFileNamingAndTriggeringPolicy>
22 |         </rollingPolicy>
23 |     </appender>
24 | 
25 |     <root level="INFO">
26 |         <appender-ref ref="STDOUT" />
27 |         <appender-ref ref="ROLLING_FILE" />
28 |     </root>
29 | 
30 |     <logger name="io.smartcat" level="DEBUG">
31 |         <appender-ref ref="STDOUT" />
32 |         <appender-ref ref="ROLLING_FILE" />
33 |     </logger>
34 | </configuration>
35 | 


--------------------------------------------------------------------------------
/cassandra-trigger/README.md:
--------------------------------------------------------------------------------
 1 | # Cassandra Trigger
 2 | 
 3 | Purpose of this project is to serve as an example for how to implement triggers in cassandra.
 4 | 
 5 | ## Create a JAR
 6 | 
 7 | This is a common Maven project with shade plugin to include all dependencies into JAR. To build it, just run:
 8 | 
 9 | `mvn clean install`
10 | 
11 | ## Use JAR in Cassandra
12 | 
13 | In order to create trigger in cassandra, JAR file needs to be places under `$CASSANDRA_CONFIG/triggers` directory on every node which will be used as coordinator. Also, path to `KafkaTrigger.yml` (line 37) needs to be adjusted to location where actuall `KafkaTrigger.yml` file is placed. Content of the file should be:
14 | 
15 | ```
16 | bootstrap.servers: cluster_kafka_1:9092,cluster_kafka_2:9092
17 | topic.name: trigger-topic
18 | ```
19 | 
20 | Note that content matches infrastcurture setup which is created using `docker-compose` command from `cluster` directory. Docker compose file used is:
21 | 
22 | ```
23 | version: '3.3'
24 | services:
25 |   zookeeper:
26 |     image: wurstmeister/zookeeper:3.4.6
27 |     ports:
28 |       - "2181:2181"
29 |   kafka:
30 |     image: wurstmeister/kafka:0.10.1.1
31 |     ports:
32 |       - 9092
33 |     environment:
34 |       HOSTNAME_COMMAND: "ifconfig | awk '/Bcast:.+/{print $$2}' | awk -F\":\" '{print $$2}'"
35 |       KAFKA_ADVERTISED_PORT: 9092
36 |       KAFKA_ZOOKEEPER_CONNECT: zookeeper:2181
37 |   cassandra-seed:
38 |     image: trigger-cassandra
39 |     ports:
40 |       - 7199
41 |       - 9042
42 |     environment:
43 |       CASSANDRA_CLUSTER_NAME: test-cluster
44 |   cassandra:
45 |     image: trigger-cassandra
46 |     ports:
47 |       - 7199
48 |       - 9042
49 |     environment:
50 |       CASSANDRA_CLUSTER_NAME: test-cluster
51 |       CASSANDRA_SEEDS: cassandra-seed
52 | ```
53 | 
54 | `trigger-cassandra` docker image is custom built for this usage. It includes `KafkaTrigger.yml` and JAR file:
55 | 
56 | ```
57 | FROM cassandra:3.11.0
58 | COPY KafkaTrigger.yml /etc/cassandra/triggers/KafkaTrigger.yml
59 | COPY cassandra-trigger-0.0.1-SNAPSHOT.jar /etc/cassandra/triggers/trigger.jar
60 | CMD ["cassandra", "-f"]
61 | ```
62 | 
63 | If you intend using JAR file in different infrastructure setup (virtual machines, different docker setup, cloud environment) configuration needs to be changed to match that infrastructure.
64 | 
65 | ## Create a trigger
66 | 
67 | To add a trigger to a table, just execute `CREATE TRIGGER kafka_trigger ON movies_by_genre USING 'io.smartcat.cassandra.trigger.KafkaTrigger';`.
68 | 
69 | 


--------------------------------------------------------------------------------
/cassandra-trigger/pom.xml:
--------------------------------------------------------------------------------
  1 | <?xml version="1.0" encoding="UTF-8"?>
  2 | <project xmlns="http://maven.apache.org/POM/4.0.0" xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance"
  3 |     xsi:schemaLocation="http://maven.apache.org/POM/4.0.0 http://maven.apache.org/xsd/maven-4.0.0.xsd">
  4 |     <modelVersion>4.0.0</modelVersion>
  5 | 
  6 |     <groupId>io.smartcat</groupId>
  7 |     <artifactId>cassandra-trigger</artifactId>
  8 |     <version>0.0.1-SNAPSHOT</version>
  9 |     <packaging>jar</packaging>
 10 | 
 11 |     <name>CassandraTrigger</name>
 12 |     <description>Custom Cassandra trigger</description>
 13 | 
 14 | 
 15 |     <properties>
 16 |         <project.build.sourceEncoding>UTF-8</project.build.sourceEncoding>
 17 |         <project.reporting.outputEncoding>UTF-8</project.reporting.outputEncoding>
 18 |         <source.level>1.8</source.level>
 19 |         <code.level>1.8</code.level>
 20 |         <version.cassandra-all>3.11.0</version.cassandra-all>
 21 |         <version.kafka>0.10.2.0</version.kafka>
 22 |         <version.plugin.compiler>3.5.1</version.plugin.compiler>
 23 |         <version.plugin.install>2.5.2</version.plugin.install>
 24 |         <version.plugin.deploy>2.8.2</version.plugin.deploy>
 25 |         <version.plugin.shade>2.4.3</version.plugin.shade>
 26 |         <version.plugin.jar>2.6</version.plugin.jar>
 27 |         <version.plugin.resources>2.7</version.plugin.resources>
 28 |     </properties>
 29 | 
 30 |     <dependencies>
 31 |         <dependency>
 32 |             <groupId>org.apache.cassandra</groupId>
 33 |             <artifactId>cassandra-all</artifactId>
 34 |             <version>${version.cassandra-all}</version>
 35 |         </dependency>
 36 |         <dependency>
 37 |             <groupId>org.apache.kafka</groupId>
 38 |             <artifactId>connect-api</artifactId>
 39 |             <version>${version.kafka}</version>
 40 |         </dependency>
 41 |         <dependency>
 42 |             <groupId>org.apache.kafka</groupId>
 43 |             <artifactId>connect-runtime</artifactId>
 44 |             <version>${version.kafka}</version>
 45 |         </dependency>
 46 |     </dependencies>
 47 | 
 48 |     <build>
 49 |         <pluginManagement>
 50 |             <plugins>
 51 |                 <plugin>
 52 |                     <groupId>org.apache.maven.plugins</groupId>
 53 |                     <artifactId>maven-jar-plugin</artifactId>
 54 |                     <version>${version.plugin.jar}</version>
 55 |                 </plugin>
 56 |                 <plugin>
 57 |                     <groupId>org.apache.maven.plugins</groupId>
 58 |                     <artifactId>maven-resources-plugin</artifactId>
 59 |                     <version>${version.plugin.resources}</version>
 60 |                 </plugin>
 61 | 
 62 |                 <plugin>
 63 |                     <groupId>org.apache.maven.plugins</groupId>
 64 |                     <artifactId>maven-install-plugin</artifactId>
 65 |                     <version>${version.plugin.install}</version>
 66 |                 </plugin>
 67 |                 <plugin>
 68 |                     <groupId>org.apache.maven.plugins</groupId>
 69 |                     <artifactId>maven-shade-plugin</artifactId>
 70 |                     <version>${version.plugin.shade}</version>
 71 |                 </plugin>
 72 |                 <plugin>
 73 |                     <groupId>org.apache.maven.plugins</groupId>
 74 |                     <artifactId>maven-deploy-plugin</artifactId>
 75 |                     <version>${version.plugin.deploy}</version>
 76 |                     <configuration>
 77 |                         <updateReleaseInfo>true</updateReleaseInfo>
 78 |                     </configuration>
 79 |                 </plugin>
 80 |                 <plugin>
 81 |                     <groupId>org.apache.maven.plugins</groupId>
 82 |                     <artifactId>maven-compiler-plugin</artifactId>
 83 |                     <version>${version.plugin.compiler}</version>
 84 |                     <inherited>true</inherited>
 85 |                     <configuration>
 86 |                         <source>${source.level}</source>
 87 |                         <target>${code.level}</target>
 88 |                         <encoding>${project.build.sourceEncoding}</encoding>
 89 |                     </configuration>
 90 |                 </plugin>
 91 |             </plugins>
 92 |         </pluginManagement>
 93 |         <plugins>
 94 |             <plugin>
 95 |                 <groupId>org.apache.maven.plugins</groupId>
 96 |                 <artifactId>maven-shade-plugin</artifactId>
 97 |                 <executions>
 98 |                     <execution>
 99 |                         <phase>package</phase>
100 |                         <goals>
101 |                             <goal>shade</goal>
102 |                         </goals>
103 |                         <configuration>
104 |                             <shadedArtifactAttached>false</shadedArtifactAttached>
105 |                             <createDependencyReducedPom>false</createDependencyReducedPom>
106 |                             <artifactSet>
107 |                                 <excludes>
108 |                                   <exclude>classworlds:classworlds</exclude>
109 |                                   <exclude>junit:junit</exclude>
110 |                                   <exclude>org.apache.maven:lib:tests</exclude>
111 |                                 </excludes>
112 |                             </artifactSet>
113 |                         </configuration>
114 |                     </execution>
115 |                 </executions>
116 |             </plugin>
117 |         </plugins>
118 |     </build>
119 | </project>
120 | 


--------------------------------------------------------------------------------
/cassandra-trigger/src/main/java/io/smartcat/cassandra/trigger/KafkaTrigger.java:
--------------------------------------------------------------------------------
  1 | package io.smartcat.cassandra.trigger;
  2 | 
  3 | import java.io.File;
  4 | import java.io.FileInputStream;
  5 | import java.io.InputStream;
  6 | import java.util.ArrayList;
  7 | import java.util.Collection;
  8 | import java.util.Collections;
  9 | import java.util.Iterator;
 10 | import java.util.List;
 11 | import java.util.Map;
 12 | import java.util.concurrent.LinkedBlockingDeque;
 13 | import java.util.concurrent.ThreadPoolExecutor;
 14 | import java.util.concurrent.TimeUnit;
 15 | 
 16 | import org.apache.cassandra.config.ColumnDefinition;
 17 | import org.apache.cassandra.db.Clustering;
 18 | import org.apache.cassandra.db.ClusteringBound;
 19 | import org.apache.cassandra.db.ClusteringPrefix;
 20 | import org.apache.cassandra.db.Mutation;
 21 | import org.apache.cassandra.db.partitions.Partition;
 22 | import org.apache.cassandra.db.rows.Cell;
 23 | import org.apache.cassandra.db.rows.Row;
 24 | import org.apache.cassandra.db.rows.Unfiltered;
 25 | import org.apache.cassandra.db.rows.UnfilteredRowIterator;
 26 | import org.apache.cassandra.io.util.FileUtils;
 27 | import org.apache.cassandra.triggers.ITrigger;
 28 | import org.apache.kafka.clients.producer.KafkaProducer;
 29 | import org.apache.kafka.clients.producer.Producer;
 30 | import org.apache.kafka.clients.producer.ProducerRecord;
 31 | import org.apache.kafka.common.serialization.StringSerializer;
 32 | import org.json.simple.JSONObject;
 33 | import org.yaml.snakeyaml.Yaml;
 34 | 
 35 | public class KafkaTrigger implements ITrigger {
 36 | 
 37 |     private static final String FILE_PATH = "/etc/cassandra/triggers/KafkaTrigger.yml";
 38 |     private static final String TOPIC_NAME = "topic.name";
 39 | 
 40 |     private final String topic;
 41 |     private final Producer<String, String> producer;
 42 |     private final ThreadPoolExecutor threadPoolExecutor;
 43 | 
 44 |     public KafkaTrigger() {
 45 |         Map<String, Object> configuration = loadConfiguration();
 46 |         topic = (String) getProperty(TOPIC_NAME, configuration);
 47 |         StringSerializer keySerializer = getSerializer(configuration, true);
 48 |         StringSerializer valueSerializer = getSerializer(configuration, false);
 49 |         producer = new KafkaProducer<>(configuration, keySerializer, valueSerializer);
 50 |         threadPoolExecutor = new ThreadPoolExecutor(4, 20, 30, TimeUnit.SECONDS, new LinkedBlockingDeque<>());
 51 |     }
 52 | 
 53 |     @Override
 54 |     public Collection<Mutation> augment(Partition partition) {
 55 |         threadPoolExecutor.execute(() -> readPartition(partition));
 56 |         return Collections.emptyList();
 57 |     }
 58 | 
 59 |     @SuppressWarnings("unchecked")
 60 |     private void readPartition(Partition partition) {
 61 |         String key = getKey(partition);
 62 |         JSONObject obj = new JSONObject();
 63 |         obj.put("key", key);
 64 |         if (partitionIsDeleted(partition)) {
 65 |             obj.put("partitionDeleted", true);
 66 |         } else {
 67 |             UnfilteredRowIterator it = partition.unfilteredIterator();
 68 |             List<JSONObject> rows = new ArrayList<>();
 69 |             while (it.hasNext()) {
 70 |                 Unfiltered un = it.next();
 71 |                 if (un.isRow()) {
 72 |                     JSONObject jsonRow = new JSONObject();
 73 |                     Clustering clustering = (Clustering) un.clustering();
 74 |                     String clusteringKey = clustering.toCQLString(partition.metadata());
 75 |                     jsonRow.put("clusteringKey", clusteringKey);
 76 |                     Row row = partition.getRow(clustering);
 77 | 
 78 |                     if (rowIsDeleted(row)) {
 79 |                         obj.put("rowDeleted", true);
 80 |                     } else {
 81 |                         Iterator<Cell> cells = row.cells().iterator();
 82 |                         Iterator<ColumnDefinition> columns = row.columns().iterator();
 83 |                         List<JSONObject> cellObjects = new ArrayList<>();
 84 |                         while (cells.hasNext() && columns.hasNext()) {
 85 |                             JSONObject jsonCell = new JSONObject();
 86 |                             ColumnDefinition columnDef = columns.next();
 87 |                             Cell cell = cells.next();
 88 |                             jsonCell.put("name", columnDef.name.toString());
 89 |                             if (cell.isTombstone()) {
 90 |                                 jsonCell.put("deleted", true);
 91 |                             } else {
 92 |                                 String data = columnDef.type.getString(cell.value());
 93 |                                 jsonCell.put("value", data);
 94 |                             }
 95 |                             cellObjects.add(jsonCell);
 96 |                         }
 97 |                         jsonRow.put("cells", cellObjects);
 98 |                     }
 99 |                     rows.add(jsonRow);
100 |                 } else if (un.isRangeTombstoneMarker()) {
101 |                     obj.put("rowRangeDeleted", true);
102 |                     ClusteringBound bound = (ClusteringBound) un.clustering();
103 |                     List<JSONObject> bounds = new ArrayList<>();
104 |                     for (int i = 0; i < bound.size(); i++) {
105 |                         String clusteringBound = partition.metadata().comparator.subtype(i).getString(bound.get(i));
106 |                         JSONObject boundObject = new JSONObject();
107 |                         boundObject.put("clusteringKey", clusteringBound);
108 |                         if (i == bound.size() - 1) {
109 |                             if (bound.kind().isStart()) {
110 |                                 boundObject.put("inclusive",
111 |                                         bound.kind() == ClusteringPrefix.Kind.INCL_START_BOUND ? true : false);
112 |                             }
113 |                             if (bound.kind().isEnd()) {
114 |                                 boundObject.put("inclusive",
115 |                                         bound.kind() == ClusteringPrefix.Kind.INCL_END_BOUND ? true : false);
116 |                             }
117 |                         }
118 |                         bounds.add(boundObject);
119 |                     }
120 |                     obj.put((bound.kind().isStart() ? "start" : "end"), bounds);
121 |                 }
122 |             }
123 |             obj.put("rows", rows);
124 |         }
125 |         String value = obj.toJSONString();
126 |         ProducerRecord<String, String> record = new ProducerRecord<>(topic, key, value);
127 |         producer.send(record);
128 |     }
129 | 
130 |     private boolean partitionIsDeleted(Partition partition) {
131 |         return partition.partitionLevelDeletion().markedForDeleteAt() > Long.MIN_VALUE;
132 |     }
133 | 
134 |     private boolean rowIsDeleted(Row row) {
135 |         return row.deletion().time().markedForDeleteAt() > Long.MIN_VALUE;
136 |     }
137 | 
138 |     private String getKey(Partition partition) {
139 |         return partition.metadata().getKeyValidator().getString(partition.partitionKey().getKey());
140 |     }
141 | 
142 |     @SuppressWarnings("unchecked")
143 |     private Map<String, Object> loadConfiguration() {
144 |         InputStream stream = null;
145 |         try {
146 |             stream = new FileInputStream(new File(FILE_PATH));
147 |             Yaml yaml = new Yaml();
148 |             return (Map<String, Object>) yaml.load(stream);
149 |         } catch (Exception e) {
150 |             throw new RuntimeException(e);
151 |         } finally {
152 |             FileUtils.closeQuietly(stream);
153 |         }
154 |     }
155 | 
156 |     private Object getProperty(String key, Map<String, Object> configuration) {
157 |         if (!configuration.containsKey(key)) {
158 |             throw new RuntimeException("Property: " + key + " not found in configuration.");
159 |         }
160 |         return configuration.get(key);
161 |     }
162 | 
163 |     private StringSerializer getSerializer(Map<String, Object> configuration, boolean isKey) {
164 |         StringSerializer serializer = new StringSerializer();
165 |         serializer.configure(configuration, isKey);
166 |         return serializer;
167 |     }
168 | }
169 | 


--------------------------------------------------------------------------------