| Property Name | Default | Description |
|---|---|---|
spark.cassandra.connection.config.cloud.path |
None | Specifies a default CloudConnectionBundle file to be for this connection. Accepts URLs (including HDFS Compatible URI's) as and references to files passed in via --files |
spark.cassandra.connection.config.profile.path |
None | Specifies a default Java Driver 4.0 Profile file to be used for this connection. Accepts URLs (including HDFS Compatible URI's) as and references to files passed in via --files |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.auth.conf.factory |
DefaultAuthConfFactory | Name of a Scala module or class implementing AuthConfFactory providing custom authentication configuration |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.connection.compression |
NONE | Compression to use (LZ4, SNAPPY or NONE) |
spark.cassandra.connection.factory |
DefaultConnectionFactory | Name of a Scala module or class implementing CassandraConnectionFactory providing connections to the Cassandra cluster |
spark.cassandra.connection.host |
localhost | Contact point to connect to the Cassandra cluster. A comma separated list may also be used. Ports may be provided but are optional. If Ports are missing spark.cassandra.connection.port will be used ("127.0.0.1:9042,192.168.0.1:9051") |
spark.cassandra.connection.keepAliveMS |
3600000 | Period of time to keep unused connections open |
spark.cassandra.connection.localConnectionsPerExecutor |
None | Number of local connections set on each Executor JVM. Defaults to the number of available CPU cores on the local node if not specified and not in a Spark Env |
spark.cassandra.connection.localDC |
None | The local DC to connect to (other nodes will be ignored) |
spark.cassandra.connection.port |
9042 | Cassandra native connection port, will be set to all hosts if no individual ports are given |
spark.cassandra.connection.quietPeriodBeforeCloseMS |
0 | The time in seconds that must pass without any additional request after requesting connection close (see Netty quiet period) |
spark.cassandra.connection.reconnectionDelayMS.max |
60000 | Maximum period of time to wait before reconnecting to a dead node |
spark.cassandra.connection.reconnectionDelayMS.min |
1000 | Minimum period of time to wait before reconnecting to a dead node |
spark.cassandra.connection.remoteConnectionsPerExecutor |
None | Minimum number of remote connections per Host set on each Executor JVM. Default value is estimated automatically based on the total number of executors in the cluster |
spark.cassandra.connection.resolveContactPoints |
true | Controls, if we need to resolve contact points at start (true), or at reconnection (false). Helpful for usage with Kubernetes or other systems with dynamic endpoints which may change while the application is running. |
spark.cassandra.connection.timeoutBeforeCloseMS |
15000 | The time in seconds for all in-flight connections to finish after requesting connection close |
spark.cassandra.connection.timeoutMS |
5000 | Maximum period of time to attempt connecting to a node |
spark.cassandra.query.retry.count |
60 | Number of times to retry a timed-out query Setting this to -1 means unlimited retries |
spark.cassandra.read.timeoutMS |
120000 | Maximum period of time to wait for a read to return |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.sql.inClauseToFullScanConversionThreshold |
20000000 | Queries with `IN` clause(s) are not converted to JoinWithCassandraTable operation if the size of cross product of all `IN` value sets exceeds this value. It is meant to stop conversion for huge `IN` values sets that may cause memory problems. If this limit is exceeded full table scan is performed. This setting takes precedence over spark.cassandra.sql.inClauseToJoinConversionThreshold. Query `select * from t where k1 in (1,2,3) and k2 in (1,2) and k3 in (1,2,3,4)` has 3 sets of `IN` values. Cross product of these values has size of 24. |
spark.cassandra.sql.inClauseToJoinConversionThreshold |
2500 | Queries with `IN` clause(s) are converted to JoinWithCassandraTable operation if the size of cross product of all `IN` value sets exceeds this value. To disable `IN` clause conversion, set this setting to 0. Query `select * from t where k1 in (1,2,3) and k2 in (1,2) and k3 in (1,2,3,4)` has 3 sets of `IN` values. Cross product of these values has size of 24. |
spark.cassandra.sql.pushdown.additionalClasses |
A comma separated list of classes to be used (in order) to apply additional pushdown rules for Cassandra Dataframes. Classes must implement CassandraPredicateRules | |
spark.cassandra.table.size.in.bytes |
None | Used by DataFrames Internally, will be updated in a future release to retrieve size from Cassandra. Can be set manually now |
spark.sql.dse.search.autoRatio |
0.03 | When Search Predicate Optimization is set to auto, Search optimizations will be preformed if this parameter * the total number of rows is greater than the number of rowsto be returned by the solr query |
spark.sql.dse.search.enableOptimization |
auto | Enables SparkSQL to automatically replace Cassandra Pushdowns with DSE Search Pushdowns utilizing lucene indexes. Valid options are On, Off, and Auto. Auto enables optimizations when the solr query will pull less than spark.sql.dse.search.autoRatio * the total table record count |
| Property Name | Default | Description |
|---|---|---|
directJoinSetting |
auto | Acceptable values, "on", "off", "auto" "on" causes a direct join to happen if possible regardless of size ratio. "off" disables direct join even when possible "auto" only does a direct join when the size ratio is satisfied see directJoinSizeRatio |
directJoinSizeRatio |
0.9 | Sets the threshold on when to perform a DirectJoin in place of a full table scan. When the size of the (CassandraSource * thisParameter) > The other side of the join, A direct join will be performed if possible. |
ignoreMissingMetaColumns |
false | Acceptable values, "true", "false" "true" ignore missing meta properties "false" throw error if missing property is requested |
ttl |
None | Surfaces the Cassandra Row TTL as a Column with the named specified. When reading use ttl.columnName=aliasForTTL. This can be done for every column with a TTL. When writing use writetime=columnName and the columname will be used to set the TTL for that row. |
writetime |
None | Surfaces the Cassandra Row Writetime as a Column with the named specified. When reading use writetime.columnName=aliasForWritetime. This can be done for every column with a writetime. When Writing use writetime=columnName and the columname will be used to set the writetime for that row. |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.connection.ssl.clientAuth.enabled |
false | Enable 2-way secure connection to Cassandra cluster |
spark.cassandra.connection.ssl.enabled |
false | Enable secure connection to Cassandra cluster |
spark.cassandra.connection.ssl.enabledAlgorithms |
Set(TLS_RSA_WITH_AES_128_CBC_SHA, TLS_RSA_WITH_AES_256_CBC_SHA) | SSL cipher suites |
spark.cassandra.connection.ssl.keyStore.password |
None | Key store password |
spark.cassandra.connection.ssl.keyStore.path |
None | Path for the key store being used |
spark.cassandra.connection.ssl.keyStore.type |
JKS | Key store type |
spark.cassandra.connection.ssl.protocol |
TLS | SSL protocol |
spark.cassandra.connection.ssl.trustStore.password |
None | Trust store password |
spark.cassandra.connection.ssl.trustStore.path |
None | Path for the trust store being used |
spark.cassandra.connection.ssl.trustStore.type |
JKS | Trust store type |
| Property Name | Default | Description |
|---|---|---|
spark.dse.continuousPagingEnabled |
true | Enables DSE Continuous Paging which improves scanning performance |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.auth.password |
None | password for password authentication |
spark.cassandra.auth.username |
None | Login name for password authentication |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.concurrent.reads |
512 | Sets read parallelism for joinWithCassandra tables |
spark.cassandra.input.consistency.level |
LOCAL_ONE | Consistency level to use when reading |
spark.cassandra.input.fetch.sizeInRows |
1000 | Number of CQL rows fetched per driver request |
spark.cassandra.input.metrics |
true | Sets whether to record connector specific metrics on write |
spark.cassandra.input.readsPerSec |
None | Sets max requests or pages per core per second, unlimited by default. |
spark.cassandra.input.split.sizeInMB |
512 | Approx amount of data to be fetched into a Spark partition. Minimum number of resulting Spark partitions is 1 + 2 * SparkContext.defaultParallelism |
spark.cassandra.input.throughputMBPerSec |
None | *(Floating points allowed)* Maximum read throughput allowed per single core in MB/s. Effects point lookups as well as full scans. |
| Property Name | Default | Description |
|---|---|---|
spark.cassandra.output.batch.grouping.buffer.size |
1000 | How many batches per single Spark task can be stored in memory before sending to Cassandra |
spark.cassandra.output.batch.grouping.key |
Partition | Determines how insert statements are grouped into batches. Available values are
|
spark.cassandra.output.batch.size.bytes |
1024 | Maximum total size of the batch in bytes. Overridden by spark.cassandra.output.batch.size.rows |
spark.cassandra.output.batch.size.rows |
None | Number of rows per single batch. The default is 'auto' which means the connector will adjust the number of rows based on the amount of data in each row |
spark.cassandra.output.concurrent.writes |
5 | Maximum number of batches executed in parallel by a single Spark task |
spark.cassandra.output.consistency.level |
LOCAL_QUORUM | Consistency level for writing |
spark.cassandra.output.ifNotExists |
false | Determines that the INSERT operation is not performed if a row with the same primary key already exists. Using the feature incurs a performance hit. |
spark.cassandra.output.ignoreNulls |
false | In Cassandra >= 2.2 null values can be left as unset in bound statements. Setting this to true will cause all null values to be left as unset rather than bound. For finer control see the CassandraOption class |
spark.cassandra.output.metrics |
true | Sets whether to record connector specific metrics on write |
spark.cassandra.output.throughputMBPerSec |
None | *(Floating points allowed)* Maximum write throughput allowed per single core in MB/s. Limit this on long (+8 hour) runs to 70% of your max throughput as seen on a smaller job for stability |
spark.cassandra.output.timestamp |
0 | Timestamp (microseconds since epoch) of the write. If not specified, the time that the write occurred is used. A value of 0 means time of write. |
spark.cassandra.output.ttl |
0 | Time To Live(TTL) assigned to writes to Cassandra. A value of 0 means no TTL |