Apache Cassandra Documentation : Cassandra Config

cluster_nameThe name of the cluster. This is mainly used to prevent machines inone logical cluster from joining another.Default Value: ‘Test Cluster’num_tokensThis defines the number of tokens randomly assigned to this node on the ringThe more tokens, relative to other nodes, the larger the proportion of datathat this node will store. You probably want all nodes to have the same numberof tokens assuming they have equal hardware capability.If you leave this unspecified, Cassandra will use the default of 1 token for legacy compatibility,and will use the initial_token as described below.Specifying initial_token will override this setting on the node’s initial start,on subsequent starts, this setting will apply even if initial token is set.If you already have a cluster with 1 token per node, and wish to migrate tomultiple tokens per node, see http://wiki.apache.org/cassandra/OperationsDefault Value: 256allocate_tokens_for_keyspaceThis option is commented out by default.Triggers automatic allocation of num_tokens tokens for this node. The allocationalgorithm attempts to choose tokens in a way that optimizes replicated load overthe nodes in the datacenter for the replication strategy used by the specifiedkeyspace.The load assigned to each node will be close to proportional to its number ofvnodes.Only supported with the Murmur3Partitioner.Default Value: KEYSPACEinitial_tokenThis option is commented out by default.initial_token allows you to specify tokens manually. While you can use it withvnodes (num_tokens > 1, above) – in which case you should provide acomma-separated list – it’s primarily used when adding nodes to legacy clustersthat do not have vnodes enabled.hinted_handoff_enabledSee http://wiki.apache.org/cassandra/HintedHandoffMay either be “true” or “false” to enable globallyDefault Value: truehinted_handoff_disabled_datacentersThis option is commented out by default.When hinted_handoff_enabled is true, a black list of data centers that will notperform hinted handoffDefault Value (complex option):# - DC1# - DC2max_hint_window_in_msthis defines the maximum amount of time a dead host will have hintsgenerated. After it has been dead this long, new hints for it will not becreated until it has been seen alive and gone down again.Default Value: 10800000 # 3 hourshinted_handoff_throttle_in_kbMaximum throttle in KBs per second, per delivery thread. This will bereduced proportionally to the number of nodes in the cluster. (If thereare two nodes in the cluster, each delivery thread will use the maximumrate; if there are three, each will throttle to half of the maximum,since we expect two nodes to be delivering hints simultaneously.)Default Value: 1024max_hints_delivery_threadsNumber of threads with which to deliver hints;Consider increasing this number when you have multi-dc deployments, sincecross-dc handoff tends to be slowerDefault Value: 2hints_directoryThis option is commented out by default.Directory where Cassandra should store hints.If not set, the default directory is $CASSANDRA_HOME/data/hints.Default Value: /var/lib/cassandra/hintshints_flush_period_in_msHow often hints should be flushed from the internal buffers to disk.Will not trigger fsync.Default Value: 10000max_hints_file_size_in_mbMaximum size for a single hints file, in megabytes.Default Value: 128hints_compressionThis option is commented out by default.Compression to apply to the hint files. If omitted, hints fileswill be written uncompressed. LZ4, Snappy, and Deflate compressorsare supported.Default Value (complex option):# - class_name: LZ4Compressor# parameters:# -batchlog_replay_throttle_in_kbMaximum throttle in KBs per second, total. This will bereduced proportionally to the number of nodes in the cluster.Default Value: 1024authenticatorAuthentication backend, implementing IAuthenticator; used to identify usersOut of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthenticator,PasswordAuthenticator}.AllowAllAuthenticator performs no checks - set it to disable authentication.PasswordAuthenticator relies on username/password pairs to authenticateusers. It keeps usernames and hashed passwords in system_auth.roles table.Please increase system_auth keyspace replication factor if you use this authenticator.If using PasswordAuthenticator, CassandraRoleManager must also be used (see below)Default Value: AllowAllAuthenticatorrole_managerPart of the Authentication & Authorization backend, implementing IRoleManager; usedto maintain grants and memberships between roles.Out of the box, Cassandra provides org.apache.cassandra.auth.CassandraRoleManager,which stores role information in the system_auth keyspace. Most functions of theIRoleManager require an authenticated login, so unless the configured IAuthenticatoractually implements authentication, most of this functionality will be unavailable.CassandraRoleManager stores role data in the system_auth keyspace. Pleaseincrease system_auth keyspace replication factor if you use this role manager.Default Value: CassandraRoleManagerroles_validity_in_msValidity period for roles cache (fetching granted roles can be an expensiveoperation depending on the role manager, CassandraRoleManager is one example)Granted roles are cached for authenticated sessions in AuthenticatedUser andafter the period specified here, become eligible for (async) reload.Defaults to 2000, set to 0 to disable caching entirely.Will be disabled automatically for AllowAllAuthenticator.Default Value: 2000permissions_validity_in_msValidity period for permissions cache (fetching permissions can be anexpensive operation depending on the authorizer, CassandraAuthorizer isone example). Defaults to 2000, set to 0 to disable.Will be disabled automatically for AllowAllAuthorizer.Default Value: 2000credentials_validity_in_msValidity period for credentials cache. This cache is tightly coupled tothe provided PasswordAuthenticator implementation of IAuthenticator. Ifanother IAuthenticator implementation is configured, this cache will notbe automatically used and so the following settings will have no effect.Please note, credentials are cached in their encrypted form, so whileactivating this cache may reduce the number of queries made to theunderlying table, it may not bring a significant reduction in thelatency of individual authentication attempts.Defaults to 2000, set to 0 to disable credentials caching.Default Value: 2000partitionerThe partitioner is responsible for distributing groups of rows (bypartition key) across nodes in the cluster. You should leave thisalone for new clusters. The partitioner can NOT be changed withoutreloading all data, so when upgrading you should set this to thesame partitioner you were already using.Besides Murmur3Partitioner, partitioners included for backwardscompatibility include RandomPartitioner, ByteOrderedPartitioner, andOrderPreservingPartitioner.Default Value: org.apache.cassandra.dht.Murmur3Partitionerdata_file_directoriesThis option is commented out by default.Directories where Cassandra should store data on disk. Cassandrawill spread data evenly across them, subject to the granularity ofthe configured compaction strategy.If not set, the default directory is $CASSANDRA_HOME/data/data.Default Value (complex option):# - /var/lib/cassandra/datacommitlog_directoryThis option is commented out by default.commit log. when running on magnetic HDD, this should be aseparate spindle than the data directories.If not set, the default directory is $CASSANDRA_HOME/data/commitlog.Default Value: /var/lib/cassandra/commitlogcdc_enabledEnable / disable CDC functionality on a per-node basis. This modifies the logic usedfor write path allocation rejection (standard: never reject. cdc: reject Mutationcontaining a CDC-enabled table if at space limit in cdc_raw_directory).Default Value: falsecdc_raw_directoryThis option is commented out by default.CommitLogSegments are moved to this directory on flush if cdc_enabled: true and thesegment contains mutations for a CDC-enabled table. This should be placed on aseparate spindle than the data directories. If not set, the default directory is$CASSANDRA_HOME/data/cdc_raw.Default Value: /var/lib/cassandra/cdc_rawdisk_failure_policyPolicy for data disk failures:dieshut down gossip and client transports and kill the JVM for any fs errors orsingle-sstable errors, so the node can be replaced.stop_paranoidshut down gossip and client transports even for single-sstable errors,kill the JVM for errors during startup.stopshut down gossip and client transports, leaving the node effectively dead, butcan still be inspected via JMX, kill the JVM for errors during startup.best_effortstop using the failed disk and respond to requests based onremaining available sstables. This means you WILL see obsoletedata at CL.ONE!ignoreignore fatal errors and let requests fail, as in pre-1.2 CassandraDefault Value: stopcommit_failure_policyPolicy for commit disk failures:dieshut down the node and kill the JVM, so the node can be replaced.stopshut down the node, leaving the node effectively dead, butcan still be inspected via JMX.stop_commitshutdown the commit log, letting writes collect butcontinuing to service reads, as in pre-2.0.5 Cassandraignoreignore fatal errors and let the batches failDefault Value: stopkey_cache_size_in_mbMaximum size of the key cache in memory.Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at theminimum, sometimes more. The key cache is fairly tiny for the amount oftime it saves, so it’s worthwhile to use it at large numbers.The row cache saves even more time, but must contain the entire row,so it is extremely space-intensive. It’s best to only use therow cache if you have hot rows or static rows.NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.Default value is empty to make it “auto” (min(5% of Heap (in MB), 100MB)). Set to 0 to disable key cache.key_cache_save_periodDuration in seconds after which Cassandra shouldsave the key cache. Caches are saved to saved_caches_directory asspecified in this configuration file.Saved caches greatly improve cold-start speeds, and is relatively cheap interms of I/O for the key cache. Row cache saving is much more expensive andhas limited use.Default is 14400 or 4 hours.Default Value: 14400key_cache_keys_to_saveThis option is commented out by default.Number of keys from the key cache to saveDisabled by default, meaning all keys are going to be savedDefault Value: 100row_cache_class_nameThis option is commented out by default.Row cache implementation class name. Available implementations:org.apache.cassandra.cache.OHCProviderFully off-heap row cache implementation (default).org.apache.cassandra.cache.SerializingCacheProviderThis is the row cache implementation availabilein previous releases of Cassandra.Default Value: org.apache.cassandra.cache.OHCProviderrow_cache_size_in_mbMaximum size of the row cache in memory.Please note that OHC cache implementation requires some additional off-heap memory to managethe map structures and some in-flight memory during operations before/after cache entries can beaccounted against the cache capacity. This overhead is usually small compared to the whole capacity.Do not specify more memory that the system can afford in the worst usual situation and leave someheadroom for OS block level cache. Do never allow your system to swap.Default value is 0, to disable row caching.Default Value: 0row_cache_save_periodDuration in seconds after which Cassandra should save the row cache.Caches are saved to saved_caches_directory as specified in this configuration file.Saved caches greatly improve cold-start speeds, and is relatively cheap interms of I/O for the key cache. Row cache saving is much more expensive andhas limited use.Default is 0 to disable saving the row cache.Default Value: 0row_cache_keys_to_saveThis option is commented out by default.Number of keys from the row cache to save.Specify 0 (which is the default), meaning all keys are going to be savedDefault Value: 100counter_cache_size_in_mbMaximum size of the counter cache in memory.Counter cache helps to reduce counter locks’ contention for hot counter cells.In case of RF = 1 a counter cache hit will cause Cassandra to skip the read beforewrite entirely. With RF > 1 a counter cache hit will still help to reduce the durationof the lock hold, helping with hot counter cell updates, but will not allow skippingthe read entirely. Only the local (clock, count) tuple of a counter cell is keptin memory, not the whole counter, so it’s relatively cheap.NOTE: if you reduce the size, you may not get you hottest keys loaded on startup.Default value is empty to make it “auto” (min(2.5% of Heap (in MB), 50MB)). Set to 0 to disable counter cache.NOTE: if you perform counter deletes and rely on low gcgs, you should disable the counter cache.counter_cache_save_periodDuration in seconds after which Cassandra shouldsave the counter cache (keys only). Caches are saved to saved_caches_directory asspecified in this configuration file.Default is 7200 or 2 hours.Default Value: 7200counter_cache_keys_to_saveThis option is commented out by default.Number of keys from the counter cache to saveDisabled by default, meaning all keys are going to be savedDefault Value: 100saved_caches_directoryThis option is commented out by default.saved cachesIf not set, the default directory is $CASSANDRA_HOME/data/saved_caches.Default Value: /var/lib/cassandra/saved_cachescommitlog_syncThis option is commented out by default.commitlog_sync may be either “periodic” or “batch.”When in batch mode, Cassandra won’t ack writes until the commit loghas been fsynced to disk. It will waitcommitlog_sync_batch_window_in_ms milliseconds between fsyncs.This window should be kept short because the writer threads willbe unable to do extra work while waiting. (You may need to increaseconcurrent_writes for the same reason.)Default Value: batchcommitlog_sync_batch_window_in_msThis option is commented out by default.Default Value: 2commitlog_syncthe other option is “periodic” where writes may be acked immediatelyand the CommitLog is simply synced every commitlog_sync_period_in_msmilliseconds.Default Value: periodiccommitlog_sync_period_in_msDefault Value: 10000commitlog_segment_size_in_mbThe size of the individual commitlog file segments. A commitlogsegment may be archived, deleted, or recycled once all the datain it (potentially from each columnfamily in the system) has beenflushed to sstables.The default size is 32, which is almost always fine, but if you arearchiving commitlog segments (see commitlog_archiving.properties),then you probably want a finer granularity of archiving; 8 or 16 MBis reasonable.Max mutation size is also configurable via max_mutation_size_in_kb setting incassandra.yaml. The default is half the size commitlog_segment_size_in_mb * 1024.NOTE: If max_mutation_size_in_kb is set explicitly then commitlog_segment_size_in_mb mustbe set to at least twice the size of max_mutation_size_in_kb / 1024Default Value: 32commitlog_compressionThis option is commented out by default.Compression to apply to the commit log. If omitted, the commit logwill be written uncompressed. LZ4, Snappy, and Deflate compressorsare supported.Default Value (complex option):# - class_name: LZ4Compressor# parameters:# -seed_providerany class that implements the SeedProvider interface and has aconstructor that takes a Map<String, String> of parameters will do.Default Value (complex option):# Addresses of hosts that are deemed contact points.# Cassandra nodes use this list of hosts to find each other and learn# the topology of the ring. You must change this if you are running# multiple nodes!- class_name: org.apache.cassandra.locator.SimpleSeedProvider parameters: # seeds is actually a comma-delimited list of addresses. # Ex: "<ip1>,<ip2>,<ip3>" - seeds: "127.0.0.1"concurrent_readsFor workloads with more data than can fit in memory, Cassandra’sbottleneck will be reads that need to fetch data fromdisk. “concurrent_reads” should be set to (16 * number_of_drives) inorder to allow the operations to enqueue low enough in the stackthat the OS and drives can reorder them. Same applies to“concurrent_counter_writes”, since counter writes read the currentvalues before incrementing and writing them back.On the other hand, since writes are almost never IO bound, the idealnumber of “concurrent_writes” is dependent on the number of cores inyour system; (8 * number_of_cores) is a good rule of thumb.Default Value: 32concurrent_writesDefault Value: 32concurrent_counter_writesDefault Value: 32concurrent_materialized_view_writesFor materialized view writes, as there is a read involved, so this shouldbe limited by the less of concurrent reads or concurrent writes.Default Value: 32file_cache_size_in_mbThis option is commented out by default.Maximum memory to use for sstable chunk cache and buffer pooling.32MB of this are reserved for pooling buffers, the rest is used as ancache that holds uncompressed sstable chunks.Defaults to the smaller of 1/4 of heap or 512MB. This pool is allocated off-heap,so is in addition to the memory allocated for heap. The cache also has on-heapoverhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved sizeif the default 64k chunk size is used).Memory is only allocated when needed.Default Value: 512buffer_pool_use_heap_if_exhaustedThis option is commented out by default.Flag indicating whether to allocate on or off heap when the sstable bufferpool is exhausted, that is when it has exceeded the maximum memoryfile_cache_size_in_mb, beyond which it will not cache buffers but allocate on request.Default Value: truedisk_optimization_strategyThis option is commented out by default.The strategy for optimizing disk readPossible values are:ssd (for solid state disks, the default)spinning (for spinning disks)Default Value: ssdmemtable_heap_space_in_mbThis option is commented out by default.Total permitted memory to use for memtables. Cassandra will stopaccepting writes when the limit is exceeded until a flush completes,and will trigger a flush based on memtable_cleanup_thresholdIf omitted, Cassandra will set both to 1/4 the size of the heap.Default Value: 2048memtable_offheap_space_in_mbThis option is commented out by default.Default Value: 2048memtable_cleanup_thresholdThis option is commented out by default.memtable_cleanup_threshold is deprecated. The default calculationis the only reasonable choice. See the comments on memtable_flush_writersfor more information.Ratio of occupied non-flushing memtable size to total permitted sizethat will trigger a flush of the largest memtable. Larger mct willmean larger flushes and hence less compaction, but also less concurrentflush activity which can make it difficult to keep your disks fedunder heavy write load.memtable_cleanup_threshold defaults to 1 / (memtable_flush_writers + 1)Default Value: 0.11memtable_allocation_typeSpecify the way Cassandra allocates and manages memtable memory.Options are:heap_bufferson heap nio buffersoffheap_buffersoff heap (direct) nio buffersoffheap_objectsoff heap objectsDefault Value: heap_bufferscommitlog_total_space_in_mbThis option is commented out by default.Total space to use for commit logs on disk.If space gets above this value, Cassandra will flush every dirty CFin the oldest segment and remove it. So a small total commitlog spacewill tend to cause more flush activity on less-active columnfamilies.The default value is the smaller of 8192, and 1/4 of the total spaceof the commitlog volume.Default Value: 8192memtable_flush_writersThis option is commented out by default.This sets the number of memtable flush writer threads per diskas well as the total number of memtables that can be flushed concurrently.These are generally a combination of compute and IO bound.Memtable flushing is more CPU efficient than memtable ingest and a single threadcan keep up with the ingest rate of a whole server on a single fast diskuntil it temporarily becomes IO bound under contention typically with compaction.At that point you need multiple flush threads. At some point in the futureit may become CPU bound all the time.You can tell if flushing is falling behind using the MemtablePool.BlockedOnAllocationmetric which should be 0, but will be non-zero if threads are blocked waiting on flushingto free memory.memtable_flush_writers defaults to two for a single data directory.This means that two memtables can be flushed concurrently to the single data directory.If you have multiple data directories the default is one memtable flushing at a timebut the flush will use a thread per data directory so you will get two or more writers.Two is generally enough to flush on a fast disk [array] mounted as a single data directory.Adding more flush writers will result in smaller more frequent flushes that introduce morecompaction overhead.There is a direct tradeoff between number of memtables that can be flushed concurrentlyand flush size and frequency. More is not better you just need enough flush writersto never stall waiting for flushing to free memory.Default Value: 2cdc_total_space_in_mbThis option is commented out by default.Total space to use for change-data-capture logs on disk.If space gets above this value, Cassandra will throw WriteTimeoutExceptionon Mutations including tables with CDC enabled. A CDCCompactor is responsiblefor parsing the raw CDC logs and deleting them when parsing is completed.The default value is the min of 4096 mb and 1/8th of the total spaceof the drive where cdc_raw_directory resides.Default Value: 4096cdc_free_space_check_interval_msThis option is commented out by default.When we hit our cdc_raw limit and the CDCCompactor is either running behindor experiencing backpressure, we check at the following interval to see if anynew space for cdc-tracked tables has been made available. Default to 250msDefault Value: 250index_summary_capacity_in_mbA fixed memory pool size in MB for for SSTable index summaries. If leftempty, this will default to 5% of the heap size. If the memory usage ofall index summaries exceeds this limit, SSTables with low read rates willshrink their index summaries in order to meet this limit. However, thisis a best-effort process. In extreme conditions Cassandra may need to usemore than this amount of memory.index_summary_resize_interval_in_minutesHow frequently index summaries should be resampled. This is doneperiodically to redistribute memory from the fixed-size pool to sstablesproportional their recent read rates. Setting to -1 will disable thisprocess, leaving existing index summaries at their current sampling level.Default Value: 60trickle_fsyncWhether to, when doing sequential writing, fsync() at intervals inorder to force the operating system to flush the dirtybuffers. Enable this to avoid sudden dirty buffer flushing fromimpacting read latencies. Almost always a good idea on SSDs; notnecessarily on platters.Default Value: falsetrickle_fsync_interval_in_kbDefault Value: 10240storage_portTCP port, for commands and dataFor security reasons, you should not expose this port to the internet. Firewall it if needed.Default Value: 7000ssl_storage_portSSL port, for encrypted communication. Unused unless enabled inencryption_optionsFor security reasons, you should not expose this port to the internet. Firewall it if needed.Default Value: 7001listen_addressAddress or interface to bind to and tell other Cassandra nodes to connect to.You _must_ change this if you want multiple nodes to be able to communicate!Set listen_address OR listen_interface, not both.Leaving it blank leaves it up to InetAddress.getLocalHost(). Thiswill always do the Right Thing _if_ the node is properly configured(hostname, name resolution, etc), and the Right Thing is to use theaddress associated with the hostname (it might not be).Setting listen_address to 0.0.0.0 is always wrong.Default Value: localhostlisten_interfaceThis option is commented out by default.Set listen_address OR listen_interface, not both. Interfaces must correspondto a single address, IP aliasing is not supported.Default Value: eth0listen_interface_prefer_ipv6This option is commented out by default.If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 addressyou can specify which should be chosen using listen_interface_prefer_ipv6. If false the first ipv4address will be used. If true the first ipv6 address will be used. Defaults to false preferringipv4. If there is only one address it will be selected regardless of ipv4/ipv6.Default Value: falsebroadcast_addressThis option is commented out by default.Address to broadcast to other Cassandra nodesLeaving this blank will set it to the same value as listen_addressDefault Value: 1.2.3.4listen_on_broadcast_addressThis option is commented out by default.When using multiple physical network interfaces, set thisto true to listen on broadcast_address in addition tothe listen_address, allowing nodes to communicate in bothinterfaces.Ignore this property if the network configuration automaticallyroutes between the public and private networks such as EC2.Default Value: falseinternode_authenticatorThis option is commented out by default.Internode authentication backend, implementing IInternodeAuthenticator;used to allow/disallow connections from peer nodes.Default Value: org.apache.cassandra.auth.AllowAllInternodeAuthenticatorstart_native_transportWhether to start the native transport server.The address on which the native transport is bound is defined by rpc_address.Default Value: truenative_transport_portport for the CQL native transport to listen for clients onFor security reasons, you should not expose this port to the internet. Firewall it if needed.Default Value: 9042native_transport_port_sslThis option is commented out by default.Enabling native transport encryption in client_encryption_options allows you to either useencryption for the standard port or to use a dedicated, additional port along with the unencryptedstandard native_transport_port.Enabling client encryption and keeping native_transport_port_ssl disabled will use encryptionfor native_transport_port. Setting native_transport_port_ssl to a different valuefrom native_transport_port will use encryption for native_transport_port_ssl whilekeeping native_transport_port unencrypted.Default Value: 9142native_transport_max_threadsThis option is commented out by default.The maximum threads for handling requests (note that idle threads are stoppedafter 30 seconds so there is not corresponding minimum setting).Default Value: 128native_transport_max_frame_size_in_mbThis option is commented out by default.The maximum size of allowed frame. Frame (requests) larger than this willbe rejected as invalid. The default is 256MB. If you’re changing this parameter,you may want to adjust max_value_size_in_mb accordingly.Default Value: 256native_transport_max_concurrent_connectionsThis option is commented out by default.The maximum number of concurrent client connections.The default is -1, which means unlimited.Default Value: -1native_transport_max_concurrent_connections_per_ipThis option is commented out by default.The maximum number of concurrent client connections per source ip.The default is -1, which means unlimited.Default Value: -1rpc_addressThe address or interface to bind the native transport server to.Set rpc_address OR rpc_interface, not both.Leaving rpc_address blank has the same effect as on listen_address(i.e. it will be based on the configured hostname of the node).Note that unlike listen_address, you can specify 0.0.0.0, but you must alsoset broadcast_rpc_address to a value other than 0.0.0.0.For security reasons, you should not expose this port to the internet. Firewall it if needed.Default Value: localhostrpc_interfaceThis option is commented out by default.Set rpc_address OR rpc_interface, not both. Interfaces must correspondto a single address, IP aliasing is not supported.Default Value: eth1rpc_interface_prefer_ipv6This option is commented out by default.If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 addressyou can specify which should be chosen using rpc_interface_prefer_ipv6. If false the first ipv4address will be used. If true the first ipv6 address will be used. Defaults to false preferringipv4. If there is only one address it will be selected regardless of ipv4/ipv6.Default Value: falsebroadcast_rpc_addressThis option is commented out by default.RPC address to broadcast to drivers and other Cassandra nodes. This cannotbe set to 0.0.0.0. If left blank, this will be set to the value ofrpc_address. If rpc_address is set to 0.0.0.0, broadcast_rpc_address mustbe set.Default Value: 1.2.3.4rpc_keepaliveenable or disable keepalive on rpc/native connectionsDefault Value: trueinternode_send_buff_size_in_bytesThis option is commented out by default.Uncomment to set socket buffer size for internode communicationNote that when setting this, the buffer size is limited by net.core.wmem_maxand when not setting it it is defined by net.ipv4.tcp_wmemSee also:/proc/sys/net/core/wmem_max/proc/sys/net/core/rmem_max/proc/sys/net/ipv4/tcp_wmem/proc/sys/net/ipv4/tcp_wmemand ‘man tcp’internode_recv_buff_size_in_bytesThis option is commented out by default.Uncomment to set socket buffer size for internode communicationNote that when setting this, the buffer size is limited by net.core.wmem_maxand when not setting it it is defined by net.ipv4.tcp_wmemincremental_backupsSet to true to have Cassandra create a hard link to each sstableflushed or streamed locally in a backups/ subdirectory of thekeyspace data. Removing these links is the operator’sresponsibility.Default Value: falsesnapshot_before_compactionWhether or not to take a snapshot before each compaction. Becareful using this option, since Cassandra won’t clean up thesnapshots for you. Mostly useful if you’re paranoid when thereis a data format change.Default Value: falseauto_snapshotWhether or not a snapshot is taken of the data before keyspace truncationor dropping of column families. The STRONGLY advised default of trueshould be used to provide data safety. If you set this flag to false, you willlose data on truncation or drop.Default Value: truecolumn_index_size_in_kbGranularity of the collation index of rows within a partition.Increase if your rows are large, or if you have a very largenumber of rows per partition. The competing goals are these:a smaller granularity means more index entries are generatedand looking up rows withing the partition by collation columnis fasterbut, Cassandra will keep the collation index in memory for hotrows (as part of the key cache), so a larger granularity meansyou can cache more hot rowsDefault Value: 64column_index_cache_size_in_kbPer sstable indexed key cache entries (the collation index in memorymentioned above) exceeding this size will not be held on heap.This means that only partition information is held on heap and theindex entries are read from disk.Note that this size refers to the size of theserialized index information and not the size of the partition.Default Value: 2concurrent_compactorsThis option is commented out by default.Number of simultaneous compactions to allow, NOT includingvalidation “compactions” for anti-entropy repair. Simultaneouscompactions can help preserve read performance in a mixed read/writeworkload, by mitigating the tendency of small sstables to accumulateduring a single long running compactions. The default is usuallyfine and if you experience problems with compaction running tooslowly or too fast, you should look atcompaction_throughput_mb_per_sec first.concurrent_compactors defaults to the smaller of (number of disks,number of cores), with a minimum of 2 and a maximum of 8.If your data directories are backed by SSD, you should increase thisto the number of cores.Default Value: 1compaction_throughput_mb_per_secThrottles compaction to the given total throughput across the entiresystem. The faster you insert data, the faster you need to compact inorder to keep the sstable count down, but in general, setting this to16 to 32 times the rate you are inserting data is more than sufficient.Setting this to 0 disables throttling. Note that this account for all typesof compaction, including validation compaction.Default Value: 16sstable_preemptive_open_interval_in_mbWhen compacting, the replacement sstable(s) can be opened before theyare completely written, and used in place of the prior sstables forany range that has been written. This helps to smoothly transfer readsbetween the sstables, reducing page cache churn and keeping hot rows hotDefault Value: 50cas_contention_timeout_in_msHow long a coordinator should continue to retry a CAS operationthat contends with other proposals for the same rowDefault Value: 1000streaming_keep_alive_period_in_secsThis option is commented out by default.Set keep-alive period for streamingThis node will send a keep-alive message periodically with this period.If the node does not receive a keep-alive message from the peer for2 keep-alive cycles the stream session times out and failDefault value is 300s (5 minutes), which means stalled streamtimes out in 10 minutes by defaultDefault Value: 300streaming_connections_per_hostThis option is commented out by default.Limit number of connections per host for streamingIncrease this when you notice that joins are CPU-bound rather that networkbound (for example a few nodes with big files).Default Value: 1phi_convict_thresholdThis option is commented out by default.phi value that must be reached for a host to be marked down.most users should never need to adjust this.Default Value: 8endpoint_snitchendpoint_snitch – Set this to a class that implementsIEndpointSnitch. The snitch has two functions:it teaches Cassandra enough about your network topology to routerequests efficientlyit allows Cassandra to spread replicas around your cluster to avoidcorrelated failures. It does this by grouping machines into“datacenters” and “racks.” Cassandra will do its best not to havemore than one replica on the same “rack” (which may not actuallybe a physical location)CASSANDRA WILL NOT ALLOW YOU TO SWITCH TO AN INCOMPATIBLE SNITCHONCE DATA IS INSERTED INTO THE CLUSTER. This would cause data loss.This means that if you start with the default SimpleSnitch, whichlocates every node on “rack1” in “datacenter1”, your only optionsif you need to add another datacenter are GossipingPropertyFileSnitch(and the older PFS). From there, if you want to migrate to anincompatible snitch like Ec2Snitch you can do it by adding new nodesunder Ec2Snitch (which will locate them in a new “datacenter”) anddecommissioning the old ones.Out of the box, Cassandra provides:SimpleSnitch:Treats Strategy order as proximity. This can improve cachelocality when disabling read repair. Only appropriate forsingle-datacenter deployments.GossipingPropertyFileSnitchThis should be your go-to snitch for production use. The rackand datacenter for the local node are defined incassandra-rackdc.properties and propagated to other nodes viagossip. If cassandra-topology.properties exists, it is used as afallback, allowing migration from the PropertyFileSnitch.PropertyFileSnitch:Proximity is determined by rack and data center, which areexplicitly configured in cassandra-topology.properties.Ec2Snitch:Appropriate for EC2 deployments in a single Region. Loads Regionand Availability Zone information from the EC2 API. The Region istreated as the datacenter, and the Availability Zone as the rack.Only private IPs are used, so this will not work across multipleRegions.Ec2MultiRegionSnitch:Uses public IPs as broadcast_address to allow cross-regionconnectivity. (Thus, you should set seed addresses to the publicIP as well.) You will need to open the storage_port orssl_storage_port on the public IP firewall. (For intra-Regiontraffic, Cassandra will switch to the private IP afterestablishing a connection.)RackInferringSnitch:Proximity is determined by rack and data center, which areassumed to correspond to the 3rd and 2nd octet of each node’s IPaddress, respectively. Unless this happens to match yourdeployment conventions, this is best used as an example ofwriting a custom Snitch class and is provided in that spirit.You can use a custom Snitch by setting this to the full class nameof the snitch, which will be assumed to be on your classpath.Default Value: SimpleSnitchdynamic_snitch_reset_interval_in_mscontrols how often to reset all host scores, allowing a bad host topossibly recoverDefault Value: 600000dynamic_snitch_badness_thresholdif set greater than zero and read_repair_chance is < 1.0, this will allow‘pinning’ of replicas to hosts in order to increase cache capacity.The badness threshold will control how much worse the pinned host has to bebefore the dynamic snitch will prefer other replicas over it. This isexpressed as a double which represents a percentage. Thus, a value of0.2 means Cassandra would continue to prefer the static snitch valuesuntil the pinned host was 20% worse than the fastest.Default Value: 0.1server_encryption_optionsEnable or disable inter-node encryptionJVM defaults for supported SSL socket protocols and cipher suites canbe replaced using custom encryption options. This is not recommendedunless you have policies in place that dictate certain settings, orneed to disable vulnerable ciphers or protocols in case the JVM cannotbe updated.FIPS compliant settings can be configured at JVM level and should notinvolve changing encryption settings here:https://docs.oracle.com/javase/8/docs/technotes/guides/security/jsse/FIPS.htmlNOTE No custom encryption options are enabled at the momentThe available internode options are : all, none, dc, rackIf set to dc cassandra will encrypt the traffic between the DCsIf set to rack cassandra will encrypt the traffic between the racksThe passwords used in these options must match the passwords used when generatingthe keystore and truststore. For instructions on generating these files, see:http://download.oracle.com/javase/6/docs/technotes/guides/security/jsse/JSSERefGuide.html#CreateKeystoreDefault Value (complex option):internode_encryption: nonekeystore: conf/.keystorekeystore_password: cassandratruststore: conf/.truststoretruststore_password: cassandra# More advanced defaults below:# protocol: TLS# algorithm: SunX509# store_type: JKS# 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]# require_client_auth: false# require_endpoint_verification: falseclient_encryption_optionsenable or disable client/server encryption.Default Value (complex option):enabled: false# If enabled and optional is set to true encrypted and unencrypted connections are handled.optional: falsekeystore: conf/.keystorekeystore_password: cassandra# require_client_auth: false# Set trustore and truststore_password if require_client_auth is true# truststore: conf/.truststore# truststore_password: cassandra# More advanced defaults below:# protocol: TLS# algorithm: SunX509# store_type: JKS# 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]internode_compressioninternode_compression controls whether traffic between nodes iscompressed.Can be:allall traffic is compresseddctraffic between different datacenters is compressednonenothing is compressed.Default Value: dcinter_dc_tcp_nodelayEnable or disable tcp_nodelay for inter-dc communication.Disabling it will result in larger (but fewer) network packets being sent,reducing overhead from the TCP protocol itself, at the cost of increasinglatency if you block for cross-datacenter responses.Default Value: falsetracetype_query_ttlTTL for different trace types used during logging of the repair process.Default Value: 86400tracetype_repair_ttlDefault Value: 604800transparent_data_encryption_optionsEnables encrypting data at-rest (on disk). Different key providers can be plugged in, but the default reads froma JCE-style keystore. A single keystore can hold multiple keys, but the one referenced bythe “key_alias” is the only key that will be used for encrypt opertaions; previously used keyscan still (and should!) be in the keystore and will be used on decrypt operations(to handle the case of key rotation).It is strongly recommended to download and install Java Cryptography Extension (JCE)Unlimited Strength Jurisdiction Policy Files for your version of the JDK.(current link: http://www.oracle.com/technetwork/java/javase/downloads/jce8-download-2133166.html)Currently, only the following file types are supported for transparent data encryption, althoughmore are coming in future cassandra releases: commitlog, hintsDefault Value (complex option):enabled: falsechunk_length_kb: 64cipher: AES/CBC/PKCS5Paddingkey_alias: testing:1# CBC IV length for AES needs to be 16 bytes (which is also the default size)# iv_length: 16key_provider: - class_name: org.apache.cassandra.security.JKSKeyProvider parameters: - keystore: conf/.keystore keystore_password: cassandra store_type: JCEKS key_password: cassandratombstone_warn_thresholdSAFETY THRESHOLDS #When executing a scan, within or across a partition, we need to keep thetombstones seen in memory so we can return them to the coordinator, whichwill use them to make sure other replicas also know about the deleted rows.With workloads that generate a lot of tombstones, this can cause performanceproblems and even exaust the server heap.(http://www.datastax.com/dev/blog/cassandra-anti-patterns-queues-and-queue-like-datasets)Adjust the thresholds here if you understand the dangers and want toscan more tombstones anyway. These thresholds may also be adjusted at runtimeusing the StorageService mbean.Default Value: 1000tombstone_failure_thresholdDefault Value: 100000batch_size_warn_threshold_in_kbLog WARN on any multiple-partition batch size exceeding this value. 5kb per batch by default.Caution should be taken on increasing the size of this threshold as it can lead to node instability.Default Value: 5batch_size_fail_threshold_in_kbFail any multiple-partition batch exceeding this value. 50kb (10x warn threshold) by default.Default Value: 50unlogged_batch_across_partitions_warn_thresholdLog WARN on any batches not of type LOGGED than span across more partitions than this limitDefault Value: 10compaction_large_partition_warning_threshold_mbLog a warning when compacting partitions larger than this valueDefault Value: 100gc_log_threshold_in_msThis option is commented out by default.GC Pauses greater than 200 ms will be logged at INFO levelThis threshold can be adjusted to minimize logging if necessaryDefault Value: 200gc_warn_threshold_in_msThis option is commented out by default.GC Pauses greater than gc_warn_threshold_in_ms will be logged at WARN levelAdjust the threshold based on your application throughput requirement. Setting to 0will deactivate the feature.Default Value: 1000max_value_size_in_mbThis option is commented out by default.Maximum size of any value in SSTables. Safety measure to detect SSTable corruptionearly. Any value size larger than this threshold will result into marking an SSTableas corrupted.Default Value: 256back_pressure_enabledBack-pressure settings #If enabled, the coordinator will apply the back-pressure strategy specified below to each mutationsent to replicas, with the aim of reducing pressure on overloaded replicas.Default Value: falseback_pressure_strategyThe back-pressure strategy applied.The default implementation, RateBasedBackPressure, takes three arguments:high ratio, factor, and flow type, and uses the ratio between incoming mutation responses and outgoing mutation requests.If below high ratio, outgoing mutations are rate limited according to the incoming rate decreased by the given factor;if above high ratio, the rate limiting is increased by the given factor;such factor is usually best configured between 1 and 10, use larger values for a faster recoveryat the expense of potentially more dropped mutations;the rate limiting is applied according to the flow type: if FAST, it’s rate limited at the speed of the fastest replica,if SLOW at the speed of the slowest one.New strategies can be added. Implementors need to implement org.apache.cassandra.net.BackpressureStrategy andprovide a public constructor accepting a Map<String, Object>.otc_coalescing_strategyThis option is commented out by default.Coalescing Strategies #Coalescing multiples messages turns out to significantly boost message processing throughput (think doubling or more).On bare metal, the floor for packet processing throughput is high enough that many applications won’t notice, but invirtualized environments, the point at which an application can be bound by network packet processing can besurprisingly low compared to the throughput of task processing that is possible inside a VM. It’s not that bare metaldoesn’t benefit from coalescing messages, it’s that the number of packets a bare metal network interface can processis sufficient for many applications such that no load starvation is experienced even without coalescing.There are other benefits to coalescing network messages that are harder to isolate with a simple metric like messagesper second. By coalescing multiple tasks together, a network thread can process multiple messages for the cost of onetrip to read from a socket, and all the task submission work can be done at the same time reducing context switchingand increasing cache friendliness of network message processing.See CASSANDRA-8692 for details.Strategy to use for coalescing messages in OutboundTcpConnection.Can be fixed, movingaverage, timehorizon, disabled (default).You can also specify a subclass of CoalescingStrategies.CoalescingStrategy by name.Default Value: DISABLEDotc_coalescing_window_usThis option is commented out by default.How many microseconds to wait for coalescing. For fixed strategy this is the amount of time after the firstmessage is received before it will be sent with any accompanying messages. For moving average this is themaximum amount of time that will be waited as well as the interval at which messages must arrive on averagefor coalescing to be enabled.Default Value: 200otc_coalescing_enough_coalesced_messagesThis option is commented out by default.Do not try to coalesce messages if we already got that many messages. This should be more than 2 and less than 128.Default Value: 8otc_backlog_expiration_interval_msThis option is commented out by default.How many milliseconds to wait between two expiration runs on the backlog (queue) of the OutboundTcpConnection.Expiration is done if messages are piling up in the backlog. Droppable messages are expired to free the memorytaken by expired messages. The interval should be between 0 and 1000, and in most installations the default valuewill be appropriate. A smaller value could potentially expire messages slightly sooner at the expense of more CPUtime and queue contention while iterating the backlog of messages.An interval of 0 disables any wait time, which is the behavior of former Cassandra versions.Default Value: 200ideal_consistency_levelThis option is commented out by default.Track a metric per keyspace indicating whether replication achieved the ideal consistencylevel for writes without timing out. This is different from the consistency level requested byeach write which may be lower in order to facilitate availability.Default Value: EACH_QUORUM