Skip to content
Maintained by AxonOps — production-grade documentation from engineers who operate distributed databases at scale

cassandra.yaml Reference

Comprehensive reference for cassandra.yaml configuration parameters, organized by category.

Version Information

Covers Apache Cassandra 4.0, 4.1, and 5.0. "Not set" = commented out by default. "Auto" = calculated from system resources.

Parameter Formats (4.1+)

Human-readable formats: 10s, 500ms (time); 256MiB, 1GiB (size)

Table of Contents

Cluster Configuration

cluster_name

Versions: 4.0, 4.1, 5.0

Default: 4.0: 'Test Cluster' | 4.1: 'Test Cluster' | 5.0: 'Test Cluster'

The name of the cluster. This is mainly used to prevent machines in one logical cluster from joining another.


num_tokens

Versions: 4.0, 4.1, 5.0

Default: 4.0: 16 | 4.1: 16 | 5.0: 16

This defines the number of tokens randomly assigned to this node on the ring The more tokens, relative to other nodes, the larger the proportion of data that this node will store. You probably want all nodes to have the same number of 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.

See https://cassandra.apache.org/doc/latest/getting_started/production.html#tokens for best practice information about num_tokens.


allocate_tokens_for_keyspace

Versions: 4.0, 4.1, 5.0

Default: 4.0: KEYSPACE | 4.1: KEYSPACE | 5.0: KEYSPACE

Replica factor is determined via the replication strategy used by the specified keyspace.


allocate_tokens_for_local_replication_factor

Versions: 4.0, 4.1, 5.0

Default: 4.0: 3 | 4.1: 3 | 5.0: 3

Replica factor is explicitly set, regardless of keyspace or datacenter. This is the replica factor within the datacenter, like NTS.


initial_token

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

initial_token allows you to specify tokens manually. While you can use it with vnodes (num_tokens > 1, above) -- in which case you should provide a comma-separated list -- it's primarily used when adding nodes to legacy clusters that do not have vnodes enabled.


partitioner

Versions: 4.0, 4.1, 5.0

Default: 4.0: org.apache.cassandra.dht.Murmur3Partitioner | 4.1: org.apache.cassandra.dht.Murmur3Partitioner | 5.0: org.apache.cassandra.dht.Murmur3Partitioner

The partitioner is responsible for distributing groups of rows (by partition key) across nodes in the cluster. The partitioner can NOT be changed without reloading all data. If you are adding nodes or upgrading, you should set this to the same partitioner that you are currently using.

The default partitioner is the Murmur3Partitioner. Older partitioners such as the RandomPartitioner, ByteOrderedPartitioner, and OrderPreservingPartitioner have been included for backward compatibility only. For new clusters, you should NOT change this value.


default_keyspace_rf

Versions: 4.1, 5.0

Default: 4.1: 1 | 5.0: 1

Impact on keyspace creation: If replication factor is not mentioned as part of keyspace creation, default_keyspace_rf would apply. Changing this configuration would only take effect for keyspaces created after the change, but does not impact existing keyspaces created prior to the change.

Impact on keyspace alter: When altering a keyspace from NetworkTopologyStrategy to SimpleStrategy, default_keyspace_rf is applied if rf is not explicitly mentioned.

Impact on system keyspaces: This would also apply for any system keyspaces that need replication factor. A further note about system keyspaces - system_traces and system_distributed keyspaces take RF of 2 or default, whichever is higher, and system_auth keyspace takes RF of 1 or default, whichever is higher.

Suggested value for use in production: 3


Network - Listen Addresses

listen_address

Versions: 4.0, 4.1, 5.0

Default: 4.0: localhost | 4.1: localhost | 5.0: localhost

Address 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(). This will always do the Right Thing if the node is properly configured (hostname, name resolution, etc), and the Right Thing is to use the address associated with the hostname (it might not be). If unresolvable it will fall back to InetAddress.getLoopbackAddress(), which is wrong for production systems.

Setting listen_address to 0.0.0.0 is always wrong.


listen_interface

Versions: 4.0, 4.1, 5.0

Default: 4.0: eth0 | 4.1: eth0 | 5.0: eth0

Set listen_address OR listen_interface, not both. Interfaces must correspond to a single address, IP aliasing is not supported.


listen_interface_prefer_ipv6

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address you can specify which should be chosen using listen_interface_prefer_ipv6. If false the first ipv4 address will be used. If true the first ipv6 address will be used. Defaults to false preferring ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.


broadcast_address

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1.2.3.4 | 4.1: 1.2.3.4 | 5.0: 1.2.3.4

Address to broadcast to other Cassandra nodes Leaving this blank will set it to the same value as listen_address


listen_on_broadcast_address

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

When using multiple physical network interfaces, set this to true to listen on broadcast_address in addition to the listen_address, allowing nodes to communicate in both interfaces. Ignore this property if the network configuration automatically routes between the public and private networks such as EC2.


storage_port

Versions: 4.0, 4.1, 5.0

Default: 4.0: 7000 | 4.1: 7000 | 5.0: 7000

TCP port, for commands and data For security reasons, you should not expose this port to the internet. Firewall it if needed.


ssl_storage_port

Versions: 4.0, 4.1, 5.0

Default: 4.0: 7001 | 4.1: 7001 | 5.0: 7001

SSL port, for legacy encrypted communication. This property is unused unless enabled in server_encryption_options (see below). As of cassandra 4.0, this property is deprecated as a single port can be used for either/both secure and insecure connections. For security reasons, you should not expose this port to the internet. Firewall it if needed.


Network - RPC/Client

rpc_address

Versions: 4.0, 4.1, 5.0

Default: 4.0: localhost | 4.1: localhost | 5.0: localhost

The 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 also set 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.


rpc_interface

Versions: 4.0, 4.1, 5.0

Default: 4.0: eth1 | 4.1: eth0 | 5.0: eth1

Set rpc_address OR rpc_interface, not both. Interfaces must correspond to a single address, IP aliasing is not supported.


rpc_interface_prefer_ipv6

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

If you choose to specify the interface by name and the interface has an ipv4 and an ipv6 address you can specify which should be chosen using rpc_interface_prefer_ipv6. If false the first ipv4 address will be used. If true the first ipv6 address will be used. Defaults to false preferring ipv4. If there is only one address it will be selected regardless of ipv4/ipv6.


broadcast_rpc_address

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1.2.3.4 | 4.1: 1.2.3.4 | 5.0: 1.2.3.4

RPC address to broadcast to drivers and other Cassandra nodes. This cannot be set to 0.0.0.0. If left blank, this will be set to the value of rpc_address. If rpc_address is set to 0.0.0.0, broadcast_rpc_address must be set.


rpc_keepalive

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

enable or disable keepalive on rpc/native connections


start_native_transport

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

Whether to start the native transport server. The address on which the native transport is bound is defined by rpc_address.


native_transport_port

Versions: 4.0, 4.1, 5.0

Default: 4.0: 9042 | 4.1: 9042 | 5.0: 9042

port for the CQL native transport to listen for clients on For security reasons, you should not expose this port to the internet. Firewall it if needed.


native_transport_port_ssl

Versions: 4.0, 4.1, 5.0

Default: 4.0: 9142 | 4.1: 9142 | 5.0: 9142

Dedicated SSL port for native transport encryption.

Enabling encryption in client_encryption_options allows two modes:

  • Single port mode: Keep native_transport_port_ssl disabled to use encryption on native_transport_port
  • Dual port mode: Set native_transport_port_ssl to a different value to have encrypted and unencrypted ports

Deprecated

This feature is deprecated since Cassandra 5.0 and will be removed. See NEWS.txt deprecation section.


native_transport_max_threads

Versions: 4.0, 4.1, 5.0

Default: 4.0: 128 | 4.1: 128 | 5.0: 128

The maximum threads for handling requests (note that idle threads are stopped after 30 seconds so there is not corresponding minimum setting).


native_transport_max_frame_size

Versions: 4.1, 5.0

Default: 4.1: 16MiB | 5.0: 16MiB

The maximum size of allowed frame. Frame (requests) larger than this will be rejected as invalid. The default is 16MiB. If you're changing this parameter, you may want to adjust max_value_size accordingly. This should be positive and less than 2048. Min unit: MiB


native_transport_max_frame_size_in_mb

Versions: 4.0

Default: 4.0: 256

The maximum size of allowed frame. Frame (requests) larger than this will be rejected as invalid. The default is 256MB. If you're changing this parameter, you may want to adjust max_value_size_in_mb accordingly. This should be positive and less than 2048.


native_transport_max_concurrent_connections

Versions: 4.0, 4.1, 5.0

Default: 4.0: -1 | 4.1: -1 | 5.0: -1

The maximum number of concurrent client connections. The default is -1, which means unlimited.


native_transport_max_concurrent_connections_per_ip

Versions: 4.0, 4.1, 5.0

Default: 4.0: -1 | 4.1: -1 | 5.0: -1

The maximum number of concurrent client connections per source ip. The default is -1, which means unlimited.


native_transport_allow_older_protocols

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

Controls whether Cassandra honors older, yet currently supported, protocol versions. The default is true, which means all supported protocols will be honored.


native_transport_rate_limiting_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

When enabled, limits the number of native transport requests dispatched for processing per second. Behavior once the limit has been breached depends on the value of THROW_ON_OVERLOAD specified in the STARTUP message sent by the client during connection establishment. (See section "4.1.1. STARTUP" in "CQL BINARY PROTOCOL v5".) With the THROW_ON_OVERLOAD flag enabled, messages that breach the limit are dropped, and an OverloadedException is thrown for the client to handle. When the flag is not enabled, the server will stop consuming messages from the channel/socket, putting backpressure on the client while already dispatched messages are processed.


native_transport_max_requests_per_second

Versions: 4.1, 5.0

Default: 4.1: 1000000 | 5.0: 1000000

No description available.


native_transport_idle_timeout

Versions: 4.1, 5.0

Default: 4.1: 60000ms | 5.0: 60000ms

Controls when idle client connections are closed. Idle connections are ones that had neither reads nor writes for a time period.

Clients may implement heartbeats by sending OPTIONS native protocol message after a timeout, which will reset idle timeout timer on the server side. To close idle client connections, corresponding values for heartbeat intervals have to be set on the client side.

Idle connection timeouts are disabled by default. Min unit: ms


native_transport_idle_timeout_in_ms

Versions: 4.0

Default: 4.0: 60000

Controls when idle client connections are closed. Idle connections are ones that had neither reads nor writes for a time period.

Clients may implement heartbeats by sending OPTIONS native protocol message after a timeout, which will reset idle timeout timer on the server side. To close idle client connections, corresponding values for heartbeat intervals have to be set on the client side.

Idle connection timeouts are disabled by default.


native_transport_flush_in_batches_legacy

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Use native transport TCP message coalescing. If on upgrade to 4.0 you found your throughput decreasing, and in particular you run an old kernel or have very fewer client connections, this option might be worth evaluating.


native_transport_max_auth_threads

Versions: 5.0

Default: 5.0: 4

The maximum threads for handling auth requests in a separate executor from main request executor. When set to 0, main executor for requests is used.


Network - Internode

internode_authenticator

Versions: 4.0, 4.1, 5.0

Default: 4.0: org.apache.cassandra.auth.AllowAllInternodeAuthenticator | 4.1: org.apache.cassandra.auth.AllowAllInternodeAuthenticator | 5.0: Not set

Internode authentication backend, implementing IInternodeAuthenticator; used to allow/disallow connections from peer nodes.


internode_socket_send_buffer_size

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Uncomment to set socket buffer size for internode communication Note that when setting this, the buffer size is limited by net.core.wmem_max and when not setting it it is defined by net.ipv4.tcp_wmem See also: /proc/sys/net/core/wmem_max /proc/sys/net/core/rmem_max /proc/sys/net/ipv4/tcp_wmem /proc/sys/net/ipv4/tcp_wmem and 'man tcp' Min unit: B


internode_socket_send_buffer_size_in_bytes

Versions: 4.0

Default: 4.0: Not set

Uncomment to set socket buffer size for internode communication Note that when setting this, the buffer size is limited by net.core.wmem_max and when not setting it it is defined by net.ipv4.tcp_wmem See also: /proc/sys/net/core/wmem_max /proc/sys/net/core/rmem_max /proc/sys/net/ipv4/tcp_wmem /proc/sys/net/ipv4/tcp_wmem and 'man tcp'


internode_socket_receive_buffer_size

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Uncomment to set socket buffer size for internode communication Note that when setting this, the buffer size is limited by net.core.wmem_max and when not setting it it is defined by net.ipv4.tcp_wmem Min unit: B


internode_socket_receive_buffer_size_in_bytes

Versions: 4.0

Default: 4.0: Not set

Uncomment to set socket buffer size for internode communication Note that when setting this, the buffer size is limited by net.core.wmem_max and when not setting it it is defined by net.ipv4.tcp_wmem


internode_application_send_queue_capacity

Versions: 4.1, 5.0

Default: 4.1: 4MiB | 5.0: 4MiB

Global, per-endpoint and per-connection limits imposed on messages queued for delivery to other nodes and waiting to be processed on arrival from other nodes in the cluster. These limits are applied to the on-wire size of the message being sent or received.

The basic per-link limit is consumed in isolation before any endpoint or global limit is imposed. Each node-pair has three links: urgent, small and large. So any given node may have a maximum of N3(internode_application_send_queue_capacity+internode_application_receive_queue_capacity) messages queued without any coordination between them although in practice, with token-aware routing, only RF*tokens nodes should need to communicate with significant bandwidth.

The per-endpoint limit is imposed on all messages exceeding the per-link limit, simultaneously with the global limit, on all links to or from a single node in the cluster. The global limit is imposed on all messages exceeding the per-link limit, simultaneously with the per-endpoint limit, on all links to or from any node in the cluster.

Min unit: B


internode_application_send_queue_capacity_in_bytes

Versions: 4.0

Default: 4.0: 4194304 #4MiB

Global, per-endpoint and per-connection limits imposed on messages queued for delivery to other nodes and waiting to be processed on arrival from other nodes in the cluster. These limits are applied to the on-wire size of the message being sent or received.

The basic per-link limit is consumed in isolation before any endpoint or global limit is imposed. Each node-pair has three links: urgent, small and large. So any given node may have a maximum of N3(internode_application_send_queue_capacity_in_bytes+internode_application_receive_queue_capacity_in_bytes) messages queued without any coordination between them although in practice, with token-aware routing, only RF*tokens nodes should need to communicate with significant bandwidth.

The per-endpoint limit is imposed on all messages exceeding the per-link limit, simultaneously with the global limit, on all links to or from a single node in the cluster. The global limit is imposed on all messages exceeding the per-link limit, simultaneously with the per-endpoint limit, on all links to or from any node in the cluster.


internode_application_send_queue_reserve_endpoint_capacity

Versions: 4.1, 5.0

Default: 4.1: 128MiB | 5.0: 128MiB

No description available.


internode_application_send_queue_reserve_endpoint_capacity_in_bytes

Versions: 4.0

Default: 4.0: 134217728 #128MiB

No description available.


internode_application_send_queue_reserve_global_capacity

Versions: 4.1, 5.0

Default: 4.1: 512MiB | 5.0: 512MiB

No description available.


internode_application_send_queue_reserve_global_capacity_in_bytes

Versions: 4.0

Default: 4.0: 536870912 #512MiB

No description available.


internode_application_receive_queue_capacity

Versions: 4.1, 5.0

Default: 4.1: 4MiB | 5.0: 4MiB

No description available.


internode_application_receive_queue_capacity_in_bytes

Versions: 4.0

Default: 4.0: 4194304 #4MiB

No description available.


internode_application_receive_queue_reserve_endpoint_capacity

Versions: 4.1, 5.0

Default: 4.1: 128MiB | 5.0: 128MiB

No description available.


internode_application_receive_queue_reserve_endpoint_capacity_in_bytes

Versions: 4.0

Default: 4.0: 134217728 #128MiB

No description available.


internode_application_receive_queue_reserve_global_capacity

Versions: 4.1, 5.0

Default: 4.1: 512MiB | 5.0: 512MiB

No description available.


internode_application_receive_queue_reserve_global_capacity_in_bytes

Versions: 4.0

Default: 4.0: 536870912 #512MiB

No description available.


internode_tcp_connect_timeout

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Defensive settings for protecting Cassandra from true network partitions. See (CASSANDRA-14358) for details.

The amount of time to wait for internode tcp connections to establish. Min unit: ms


internode_tcp_connect_timeout_in_ms

Versions: 4.0

Default: 4.0: 2000

Defensive settings for protecting Cassandra from true network partitions. See (CASSANDRA-14358) for details.

The amount of time to wait for internode tcp connections to establish.


internode_tcp_user_timeout

Versions: 4.1, 5.0

Default: 4.1: 30000ms | 5.0: 30000ms

The amount of time unacknowledged data is allowed on a connection before we throw out the connection Note this is only supported on Linux + epoll, and it appears to behave oddly above a setting of 30000 (it takes much longer than 30s) as of Linux 4.12. If you want something that high set this to 0 which picks up the OS default and configure the net.ipv4.tcp_retries2 sysctl to be ~8. Min unit: ms


internode_tcp_user_timeout_in_ms

Versions: 4.0

Default: 4.0: 30000

The amount of time unacknowledged data is allowed on a connection before we throw out the connection Note this is only supported on Linux + epoll, and it appears to behave oddly above a setting of 30000 (it takes much longer than 30s) as of Linux 4.12. If you want something that high set this to 0 which picks up the OS default and configure the net.ipv4.tcp_retries2 sysctl to be ~8.


internode_streaming_tcp_user_timeout

Versions: 4.1, 5.0

Default: 4.1: 300000ms | 5.0: 300000ms

The amount of time unacknowledged data is allowed on a streaming connection. The default is 5 minutes. Increase it or set it to 0 in order to increase the timeout. Min unit: ms


internode_streaming_tcp_user_timeout_in_ms

Versions: 4.0

Default: 4.0: 300000

The amount of time unacknowledged data is allowed on a streaming connection. The default is 5 minutes. Increase it or set it to 0 in order to increase the timeout.


internode_compression

Versions: 4.0, 4.1, 5.0

Default: 4.0: dc | 4.1: dc | 5.0: dc

internode_compression controls whether traffic between nodes is compressed. Can be:

all all traffic is compressed

dc traffic between different datacenters is compressed

none nothing is compressed.


cross_node_timeout

Versions: 4.0

Default: 4.0: true

Enable operation timeout information exchange between nodes to accurately measure request timeouts. If disabled, replicas will assume that requests were forwarded to them instantly by the coordinator, which means that under overload conditions we will waste that much extra time processing already-timed-out requests.

Warning: It is generally assumed that users have setup NTP on their clusters, and that clocks are modestly in sync, since this is a requirement for general correctness of last write wins.


inter_dc_tcp_nodelay

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Enable 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 increasing latency if you block for cross-datacenter responses.


internode_timeout

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Enable operation timeout information exchange between nodes to accurately measure request timeouts. If disabled, replicas will assume that requests were forwarded to them instantly by the coordinator, which means that under overload conditions we will waste that much extra time processing already-timed-out requests.

Warning: It is generally assumed that users have setup NTP on their clusters, and that clocks are modestly in sync, since this is a requirement for general correctness of last write wins.


Seed Provider

seed_provider

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Configures how the node discovers other nodes in the cluster.

seed_provider:
  - class_name: org.apache.cassandra.locator.SimpleSeedProvider
    parameters:
      - seeds: "127.0.0.1:7000"
Sub-parameter Description
class_name The seed provider class. Default: SimpleSeedProvider
seeds Comma-separated list of seed node addresses (host:port)

Data Directories

data_file_directories

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

Directories where Cassandra stores SSTable data files.

data_file_directories:
  - /var/lib/cassandra/data

This is a list of directory paths. Cassandra will spread data evenly across all specified directories.


commitlog_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: /var/lib/cassandra/commitlog | 4.1: /var/lib/cassandra/commitlog | 5.0: /var/lib/cassandra/commitlog

commit log. when running on magnetic HDD, this should be a separate spindle than the data directories. If not set, the default directory is $CASSANDRA_HOME/data/commitlog.


saved_caches_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: /var/lib/cassandra/saved_caches | 4.1: /var/lib/cassandra/saved_caches | 5.0: /var/lib/cassandra/saved_caches

saved caches If not set, the default directory is $CASSANDRA_HOME/data/saved_caches.


hints_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: /var/lib/cassandra/hints | 4.1: /var/lib/cassandra/hints | 5.0: /var/lib/cassandra/hints

Directory where Cassandra should store hints. If not set, the default directory is $CASSANDRA_HOME/data/hints.


cdc_raw_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: /var/lib/cassandra/cdc_raw | 4.1: /var/lib/cassandra/cdc_raw | 5.0: /var/lib/cassandra/cdc_raw

CommitLogSegments are moved to this directory on flush if cdc_enabled: true and the segment contains mutations for a CDC-enabled table. This should be placed on a separate spindle than the data directories. If not set, the default directory is $CASSANDRA_HOME/data/cdc_raw.


local_system_data_file_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

Directory were Cassandra should store the data of the local system keyspaces. By default Cassandra will store the data of the local system keyspaces in the first of the data directories specified by data_file_directories. This approach ensures that if one of the other disks is lost Cassandra can continue to operate. For extra security this setting allows to store those data on a different directory that provides redundancy.


Hinted Handoff

hinted_handoff_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

May either be "true" or "false" to enable globally


max_hint_window

Versions: 4.1, 5.0

Default: 4.1: 3h | 5.0: 3h

this defines the maximum amount of time a dead host will have hints generated. After it has been dead this long, new hints for it will not be created until it has been seen alive and gone down again. Min unit: ms


max_hint_window_in_ms

Versions: 4.0

Default: 4.0: 10800000

this defines the maximum amount of time a dead host will have hints generated. After it has been dead this long, new hints for it will not be created until it has been seen alive and gone down again.


hinted_handoff_disabled_datacenters

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

When hinted_handoff_enabled is true, a black list of data centers that will not perform hinted handoff


max_hints_delivery_threads

Versions: 4.0, 4.1, 5.0

Default: 4.0: 2 | 4.1: 2 | 5.0: 2

Number of threads with which to deliver hints; Consider increasing this number when you have multi-dc deployments, since cross-dc handoff tends to be slower


hints_flush_period

Versions: 4.1, 5.0

Default: 4.1: 10000ms | 5.0: 10000ms

How often hints should be flushed from the internal buffers to disk. Will not trigger fsync. Min unit: ms


hints_flush_period_in_ms

Versions: 4.0

Default: 4.0: 10000

How often hints should be flushed from the internal buffers to disk. Will not trigger fsync.


max_hints_file_size

Versions: 4.1, 5.0

Default: 4.1: 128MiB | 5.0: 128MiB

Maximum size for a single hints file, in mebibytes. Min unit: MiB


max_hints_file_size_in_mb

Versions: 4.0

Default: 4.0: 128

Maximum size for a single hints file, in megabytes.


hint_window_persistent_enabled

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Enable / disable persistent hint windows.

If set to false, a hint will be stored only in case a respective node that hint is for is down less than or equal to max_hint_window.

If set to true, a hint will be stored in case there is not any hint which was stored earlier than max_hint_window. This is for cases when a node keeps to restart and hints are not delivered yet, we would be saving hints for that node indefinitely.

Defaults to true.


auto_hints_cleanup_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Enable / disable automatic cleanup for the expired and orphaned hints file. Disable the option in order to preserve those hints on the disk.


hinted_handoff_throttle

Versions: 4.1, 5.0

Default: 4.1: 1024KiB | 5.0: 1024KiB

Maximum throttle in KiBs per second, per delivery thread. This will be reduced proportionally to the number of nodes in the cluster. (If there are two nodes in the cluster, each delivery thread will use the maximum rate; if there are three, each will throttle to half of the maximum, since we expect two nodes to be delivering hints simultaneously.) Min unit: KiB


hinted_handoff_throttle_in_kb

Versions: 4.0

Default: 4.0: 1024

Maximum throttle in KBs per second, per delivery thread. This will be reduced proportionally to the number of nodes in the cluster. (If there are two nodes in the cluster, each delivery thread will use the maximum rate; if there are three, each will throttle to half of the maximum, since we expect two nodes to be delivering hints simultaneously.)


hints_compression

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

Configures compression for hint files.

hints_compression:
  - class_name: LZ4Compressor

Same compression options as commitlog_compression.


transfer_hints_on_decommission

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Enable/disable transfering hints to a peer during decommission. Even when enabled, this does not guarantee consistency for logged batches, and it may delay decommission when coupled with a strict hinted_handoff_throttle. Default: true


max_hints_size_per_host

Versions: 4.1, 5.0

Default: 4.1: 0MiB | 5.0: 0MiB

The file size limit to store hints for an unreachable host, in mebibytes. Once the local hints files have reached the limit, no more new hints will be created. Set a non-positive value will disable the size limit.


Memtable Configuration

memtable_allocation_type

Versions: 4.0, 4.1, 5.0

Default: 4.0: heap_buffers | 4.1: heap_buffers | 5.0: heap_buffers

Specify the way Cassandra allocates and manages memtable memory. Options are:

heap_buffers on heap nio buffers

offheap_buffers off heap (direct) nio buffers

offheap_objects off heap objects


memtable_heap_space

Versions: 4.1, 5.0

Default: 4.1: 2048MiB | 5.0: 2048MiB

Total permitted memory to use for memtables. Cassandra will stop accepting writes when the limit is exceeded until a flush completes, and will trigger a flush based on memtable_cleanup_threshold If omitted, Cassandra will set both to 1/4 the size of the heap. Min unit: MiB


memtable_heap_space_in_mb

Versions: 4.0

Default: 4.0: 2048

Total permitted memory to use for memtables. Cassandra will stop accepting writes when the limit is exceeded until a flush completes, and will trigger a flush based on memtable_cleanup_threshold If omitted, Cassandra will set both to 1/4 the size of the heap.


memtable_offheap_space

Versions: 4.1, 5.0

Default: 4.1: 2048MiB | 5.0: 2048MiB

Min unit: MiB


memtable_offheap_space_in_mb

Versions: 4.0

Default: 4.0: 2048

No description available.


memtable_cleanup_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 0.11 | 4.1: 0.11 | 5.0: 0.11

memtable_cleanup_threshold is deprecated. The default calculation is the only reasonable choice. See the comments on memtable_flush_writers for more information.

Ratio of occupied non-flushing memtable size to total permitted size that will trigger a flush of the largest memtable. Larger mct will mean larger flushes and hence less compaction, but also less concurrent flush activity which can make it difficult to keep your disks fed under heavy write load.

memtable_cleanup_threshold defaults to 1 / (memtable_flush_writers + 1)


memtable_flush_writers

Versions: 4.0, 4.1, 5.0

Default: 4.0: 2 | 4.1: 2 | 5.0: 2

This sets the number of memtable flush writer threads per disk as 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 thread can keep up with the ingest rate of a whole server on a single fast disk until it temporarily becomes IO bound under contention typically with compaction. At that point you need multiple flush threads. At some point in the future it may become CPU bound all the time.

You can tell if flushing is falling behind using the MemtablePool.BlockedOnAllocation metric which should be 0, but will be non-zero if threads are blocked waiting on flushing to 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 time but 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 more compaction overhead.

There is a direct tradeoff between number of memtables that can be flushed concurrently and flush size and frequency. More is not better you just need enough flush writers to never stall waiting for flushing to free memory.


flush_compression

Versions: 4.0, 4.1, 5.0

Default: 4.0: fast | 4.1: fast | 5.0: fast

was the pre 4.0 behavior.


memtable

Versions: 5.0

Default: 5.0: (see below)

Configures memtable implementation.

memtable:
  configurations:
    default:
      class_name: TrieMemtable

Implementations: SkipListMemtable (legacy), TrieMemtable (reduced GC, better throughput)


Commit Log

commitlog_sync

Versions: 4.0, 4.1, 5.0

Default: 4.0: periodic | 4.1: periodic | 5.0: periodic

the default option is "periodic" where writes may be acked immediately and the CommitLog is simply synced every commitlog_sync_period_in_ms milliseconds.


commitlog_sync_batch_window_in_ms

Versions: 4.0, 4.1

Default: 4.0: 2 | 4.1: 2

commitlog_sync may be either "periodic", "group", or "batch."

When in batch mode, Cassandra won't ack writes until the commit log has been flushed to disk. Each incoming write will trigger the flush task. commitlog_sync_batch_window_in_ms is a deprecated value. Previously it had almost no value, and is being removed.


commitlog_sync_group_window

Versions: 4.1, 5.0

Default: 4.1: 1000ms | 5.0: 1000ms

commitlog_sync may be either "periodic", "group", or "batch."

When in batch mode, Cassandra won't ack writes until the commit log has been flushed to disk. Each incoming write will trigger the flush task.

group mode is similar to batch mode, where Cassandra will not ack writes until the commit log has been flushed to disk. The difference is group mode will wait up to commitlog_sync_group_window between flushes.

Min unit: ms


commitlog_sync_group_window_in_ms

Versions: 4.0

Default: 4.0: 1000

group mode is similar to batch mode, where Cassandra will not ack writes until the commit log has been flushed to disk. The difference is group mode will wait up to commitlog_sync_group_window_in_ms between flushes.


commitlog_sync_period

Versions: 4.1, 5.0

Default: 4.1: 10000ms | 5.0: 10000ms

Min unit: ms


commitlog_sync_period_in_ms

Versions: 4.0

Default: 4.0: 10000

No description available.


commitlog_segment_size

Versions: 4.1, 5.0

Default: 4.1: 32MiB | 5.0: 32MiB

The size of the individual commitlog file segments. A commitlog segment may be archived, deleted, or recycled once all the data in it (potentially from each columnfamily in the system) has been flushed to sstables.

The default size is 32, which is almost always fine, but if you are archiving commitlog segments (see commitlog_archiving.properties), then you probably want a finer granularity of archiving; 8 or 16 MB is reasonable. Max mutation size is also configurable via max_mutation_size setting in cassandra.yaml. The default is half the size commitlog_segment_size in bytes. This should be positive and less than 2048.

NOTE: If max_mutation_size is set explicitly then commitlog_segment_size must be set to at least twice the size of max_mutation_size

Min unit: MiB


commitlog_segment_size_in_mb

Versions: 4.0

Default: 4.0: 32

The size of the individual commitlog file segments. A commitlog segment may be archived, deleted, or recycled once all the data in it (potentially from each columnfamily in the system) has been flushed to sstables.

The default size is 32, which is almost always fine, but if you are archiving commitlog segments (see commitlog_archiving.properties), then you probably want a finer granularity of archiving; 8 or 16 MB is reasonable. Max mutation size is also configurable via max_mutation_size_in_kb setting in cassandra.yaml. The default is half the size commitlog_segment_size_in_mb * 1024. This should be positive and less than 2048.

NOTE: If max_mutation_size_in_kb is set explicitly then commitlog_segment_size_in_mb must be set to at least twice the size of max_mutation_size_in_kb / 1024


commitlog_compression

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

Configures compression for commit log segments.

commitlog_compression:
  - class_name: LZ4Compressor
    parameters:
Sub-parameter Description
class_name LZ4Compressor, SnappyCompressor, DeflateCompressor, or ZstdCompressor
parameters Algorithm-specific parameters (optional)

commitlog_total_space

Versions: 4.1, 5.0

Default: 4.1: 8192MiB | 5.0: 8192MiB

Total space to use for commit logs on disk.

If space gets above this value, Cassandra will flush every dirty CF in the oldest segment and remove it. So a small total commitlog space will tend to cause more flush activity on less-active columnfamilies.

The default value is the smaller of 8192, and 1/4 of the total space of the commitlog volume.


commitlog_total_space_in_mb

Versions: 4.0

Default: 4.0: 8192

Total space to use for commit logs on disk.

If space gets above this value, Cassandra will flush every dirty CF in the oldest segment and remove it. So a small total commitlog space will tend to cause more flush activity on less-active columnfamilies.

The default value is the smaller of 8192, and 1/4 of the total space of the commitlog volume.


commitlog_disk_access_mode

Versions: 5.0

Default: 5.0: legacy

Set the disk access mode for writing commitlog segments. The allowed values are: - auto: version dependent optimal setting - legacy: the default mode as used in Cassandra 4.x and earlier (standard I/O when the commitlog is either compressed or encrypted or mmap otherwise) - mmap: use memory mapped I/O - available only when the commitlog is neither compressed nor encrypted - direct: use direct I/O - available only when the commitlog is neither compressed nor encrypted - standard: use standard I/O - available only when the commitlog is compressed or encrypted The default setting is legacy when the storage compatibility is set to 4 or auto otherwise.


periodic_commitlog_sync_lag_block

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

When in periodic commitlog mode, the number of milliseconds to block writes while waiting for a slow disk flush to complete. Min unit: ms


periodic_commitlog_sync_lag_block_in_ms

Versions: 4.0

Default: 4.0: Not set

When in periodic commitlog mode, the number of milliseconds to block writes while waiting for a slow disk flush to complete.


Concurrent Operations

concurrent_reads

Versions: 4.0, 4.1, 5.0

Default: 4.0: 32 | 4.1: 32 | 5.0: 32

For workloads with more data than can fit in memory, Cassandra's bottleneck will be reads that need to fetch data from disk. "concurrent_reads" should be set to (16 * number_of_drives) in order to allow the operations to enqueue low enough in the stack that the OS and drives can reorder them. Same applies to "concurrent_counter_writes", since counter writes read the current values before incrementing and writing them back.

On the other hand, since writes are almost never IO bound, the ideal number of "concurrent_writes" is dependent on the number of cores in your system; (8 * number_of_cores) is a good rule of thumb.


concurrent_writes

Versions: 4.0, 4.1, 5.0

Default: 4.0: 32 | 4.1: 32 | 5.0: 32

No description available.


concurrent_counter_writes

Versions: 4.0, 4.1, 5.0

Default: 4.0: 32 | 4.1: 32 | 5.0: 32

No description available.


concurrent_materialized_view_writes

Versions: 4.0, 4.1, 5.0

Default: 4.0: 32 | 4.1: 32 | 5.0: 32

For materialized view writes, as there is a read involved, so this should be limited by the less of concurrent reads or concurrent writes.


concurrent_compactors

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1 | 4.1: 1 | 5.0: 1

Number of simultaneous compactions to allow, NOT including validation "compactions" for anti-entropy repair. Simultaneous compactions can help preserve read performance in a mixed read/write workload, by mitigating the tendency of small sstables to accumulate during a single long running compactions. The default is usually fine and if you experience problems with compaction running too slowly or too fast, you should look at compaction_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 this to the number of cores.


concurrent_validations

Versions: 4.0, 4.1, 5.0

Default: 4.0: 0 | 4.1: 0 | 5.0: 0

Number of simultaneous repair validations to allow. If not set or set to a value less than 1, it defaults to the value of concurrent_compactors. To set a value greeater than concurrent_compactors at startup, the system property cassandra.allow_unlimited_concurrent_validations must be set to true. To dynamically resize to a value > concurrent_compactors on a running node, first call the bypassConcurrentValidatorsLimit method on the org.apache.cassandra.db:type=StorageService mbean


concurrent_materialized_view_builders

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1 | 4.1: 1 | 5.0: 1

Number of simultaneous materialized view builder tasks to allow.


concurrent_merkle_tree_requests

Versions: 4.1, 5.0

Default: 4.1: 0 | 5.0: 0

The number of simultaneous Merkle tree requests during repairs that can be performed by a repair command. The size of each validation request is limited by the repair_session_space property, so setting this to 1 will make sure that a repair command doesn't exceed that limit, even if the repair command is repairing multiple tables or multiple virtual nodes.

There isn't a limit by default for backwards compatibility, but this can produce OOM for commands repairing multiple tables or multiple virtual nodes. A limit of just 1 simultaneous Merkle tree request is generally recommended with no virtual nodes so repair_session_space, and thereof the Merkle tree resolution, can be high. For virtual nodes a value of 1 with the default repair_session_space value will produce higher resolution Merkle trees at the expense of speed. Alternatively, when working with virtual nodes it can make sense to reduce the repair_session_space and increase the value of concurrent_merkle_tree_requests because each range will contain fewer data.

For more details see https://issues.apache.org/jira/browse/CASSANDRA-19336.

A zero value means no limit.


Compaction

compaction_throughput

Versions: 4.1, 5.0

Default: 4.1: 64MiB/s | 5.0: 64MiB/s

Throttles compaction to the given total throughput across the entire system. The faster you insert data, the faster you need to compact in order to keep the sstable count down, but in general, setting this to 16 to 32 times the rate you are inserting data is more than sufficient. Setting this to 0 disables throttling. Note that this accounts for all types of compaction, including validation compaction (building Merkle trees for repairs).


compaction_throughput_mb_per_sec

Versions: 4.0

Default: 4.0: 64

Throttles compaction to the given total throughput across the entire system. The faster you insert data, the faster you need to compact in order to keep the sstable count down, but in general, setting this to 16 to 32 times the rate you are inserting data is more than sufficient. Setting this to 0 disables throttling. Note that this accounts for all types of compaction, including validation compaction (building Merkle trees for repairs).


compaction_large_partition_warning_threshold

Versions: 4.1

Default: 4.1: 100MiB

Log a warning when compacting partitions larger than this value


compaction_large_partition_warning_threshold_mb

Versions: 4.0

Default: 4.0: 100

Log a warning when compacting partitions larger than this value


compaction_tombstone_warning_threshold

Versions: 4.1

Default: 4.1: 100000

Log a warning when writing more tombstones than this value to a partition


sstable_preemptive_open_interval

Versions: 4.1, 5.0

Default: 4.1: 50MiB | 5.0: 50MiB

When compacting, the replacement sstable(s) can be opened before they are completely written, and used in place of the prior sstables for any range that has been written. This helps to smoothly transfer reads between the sstables, reducing page cache churn and keeping hot rows hot Set sstable_preemptive_open_interval to null for disabled which is equivalent to sstable_preemptive_open_interval_in_mb being negative Min unit: MiB


sstable_preemptive_open_interval_in_mb

Versions: 4.0

Default: 4.0: 50

When compacting, the replacement sstable(s) can be opened before they are completely written, and used in place of the prior sstables for any range that has been written. This helps to smoothly transfer reads between the sstables, reducing page cache churn and keeping hot rows hot


uuid_sstable_identifiers_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Starting from 4.1 sstables support UUID based generation identifiers. They are disabled by default because once enabled, there is no easy way to downgrade. When the node is restarted with this option set to true, each newly created sstable will have a UUID based generation identifier and such files are not readable by previous Cassandra versions. At some point, this option will become true by default and eventually get removed from the configuration.


automatic_sstable_upgrade

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Automatically upgrade sstables after upgrade - if there is no ordinary compaction to do, the oldest non-upgraded sstable will get upgraded to the latest version


default_compaction

Versions: 5.0

Default: 5.0: Not set

Default compaction strategy applied when a table doesn't specify compaction. Also applies to system tables.

default_compaction:
  class_name: SizeTieredCompactionStrategy
  parameters:
    min_threshold: 4
    max_threshold: 32
Sub-parameter Description Default
class_name Compaction strategy class SizeTieredCompactionStrategy
min_threshold Minimum SSTables to trigger compaction 4
max_threshold Maximum SSTables to compact at once 32

If not specified, defaults to SizeTieredCompactionStrategy with default parameters.


max_concurrent_automatic_sstable_upgrades

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1 | 4.1: 1 | 5.0: 1

Limit the number of concurrent sstable upgrades


Caching

key_cache_size

Versions: 4.1, 5.0

Default: 4.1: Auto | 5.0: Auto

Maximum size of the key cache in memory.

Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at the minimum, sometimes more. The key cache is fairly tiny for the amount of time 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 the row 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 MiB), 100MiB)). Set to 0 to disable key cache.

This is only relevant to SSTable formats that use key cache, e.g. BIG. Min unit: MiB


key_cache_size_in_mb

Versions: 4.0

Default: 4.0: Auto

Maximum size of the key cache in memory.

Each key cache hit saves 1 seek and each row cache hit saves 2 seeks at the minimum, sometimes more. The key cache is fairly tiny for the amount of time 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 the row 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_period

Versions: 4.0, 4.1, 5.0

Default: 4.0: 14400 | 4.1: 4h | 5.0: 4h

Duration in seconds after which Cassandra should save the key 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 in terms of I/O for the key cache. Row cache saving is much more expensive and has limited use.

This is only relevant to SSTable formats that use key cache, e.g. BIG. Default is 14400 or 4 hours. Min unit: s


key_cache_keys_to_save

Versions: 4.0, 4.1, 5.0

Default: 4.0: 100 | 4.1: 100 | 5.0: 100

Number of keys from the key cache to save Disabled by default, meaning all keys are going to be saved This is only relevant to SSTable formats that use key cache, e.g. BIG.


row_cache_size

Versions: 4.1, 5.0

Default: 4.1: 0MiB | 5.0: 0MiB

Maximum size of the row cache in memory. Please note that OHC cache implementation requires some additional off-heap memory to manage the map structures and some in-flight memory during operations before/after cache entries can be accounted 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 some headroom for OS block level cache. Do never allow your system to swap.

Default value is 0, to disable row caching. Min unit: MiB


row_cache_size_in_mb

Versions: 4.0

Default: 4.0: 0

Maximum size of the row cache in memory. Please note that OHC cache implementation requires some additional off-heap memory to manage the map structures and some in-flight memory during operations before/after cache entries can be accounted 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 some headroom for OS block level cache. Do never allow your system to swap.

Default value is 0, to disable row caching.


row_cache_save_period

Versions: 4.0, 4.1, 5.0

Default: 4.0: 0 | 4.1: 0s | 5.0: 0s

Duration 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 in terms of I/O for the key cache. Row cache saving is much more expensive and has limited use.

Default is 0 to disable saving the row cache. Min unit: s


row_cache_keys_to_save

Versions: 4.0, 4.1, 5.0

Default: 4.0: 100 | 4.1: 100 | 5.0: 100

Number of keys from the row cache to save. Specify 0 (which is the default), meaning all keys are going to be saved


row_cache_class_name

Versions: 4.0, 4.1, 5.0

Default: 4.0: org.apache.cassandra.cache.OHCProvider | 4.1: org.apache.cassandra.cache.OHCProvider | 5.0: org.apache.cassandra.cache.OHCProvider

Row cache implementation class name. Available implementations:

org.apache.cassandra.cache.OHCProvider Fully off-heap row cache implementation (default).

org.apache.cassandra.cache.SerializingCacheProvider This is the row cache implementation availabile in previous releases of Cassandra.


counter_cache_size

Versions: 4.1, 5.0

Default: 4.1: Auto | 5.0: Auto

Maximum 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 before write entirely. With RF > 1 a counter cache hit will still help to reduce the duration of the lock hold, helping with hot counter cell updates, but will not allow skipping the read entirely. Only the local (clock, count) tuple of a counter cell is kept in 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 MiB), 50MiB)). Set to 0 to disable counter cache. NOTE: if you perform counter deletes and rely on low gcgs, you should disable the counter cache. Min unit: MiB


counter_cache_size_in_mb

Versions: 4.0

Default: 4.0: Auto

Maximum 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 before write entirely. With RF > 1 a counter cache hit will still help to reduce the duration of the lock hold, helping with hot counter cell updates, but will not allow skipping the read entirely. Only the local (clock, count) tuple of a counter cell is kept in 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_period

Versions: 4.0, 4.1, 5.0

Default: 4.0: 7200 | 4.1: 7200s | 5.0: 7200s

Duration in seconds after which Cassandra should save the counter cache (keys only). Caches are saved to saved_caches_directory as specified in this configuration file.

Default is 7200 or 2 hours. Min unit: s


counter_cache_keys_to_save

Versions: 4.0, 4.1, 5.0

Default: 4.0: 100 | 4.1: 100 | 5.0: 100

Number of keys from the counter cache to save Disabled by default, meaning all keys are going to be saved


cache_load_timeout

Versions: 4.1, 5.0

Default: 4.1: 30s | 5.0: 30s

Number of seconds the server will wait for each cache (row, key, etc ...) to load while starting the Cassandra process. Setting this to zero is equivalent to disabling all cache loading on startup while still having the cache during runtime. Min unit: s


cache_load_timeout_seconds

Versions: 4.0

Default: 4.0: 30

Number of seconds the server will wait for each cache (row, key, etc ...) to load while starting the Cassandra process. Setting this to a negative value is equivalent to disabling all cache loading on startup while still having the cache during runtime.


file_cache_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Enable the sstable chunk cache. The chunk cache will store recently accessed sections of the sstable in-memory as uncompressed buffers.


file_cache_size

Versions: 4.1, 5.0

Default: 4.1: 512MiB | 5.0: 512MiB

Maximum memory to use for sstable chunk cache and buffer pooling. 32MB of this are reserved for pooling buffers, the rest is used for chunk cache 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-heap overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size if the default 64k chunk size is used). Memory is only allocated when needed. Min unit: MiB


file_cache_size_in_mb

Versions: 4.0

Default: 4.0: 512

Maximum memory to use for sstable chunk cache and buffer pooling. 32MB of this are reserved for pooling buffers, the rest is used for chunk cache 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-heap overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size if the default 64k chunk size is used). Memory is only allocated when needed.


buffer_pool_use_heap_if_exhausted

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

No description available.


prepared_statements_cache_size

Versions: 4.1, 5.0

Default: 4.1: Auto | 5.0: Auto

Maximum size of the native protocol prepared statement cache

Valid values are either "auto" (omitting the value) or a value greater 0.

Note that specifying a too large value will result in long running GCs and possbily out-of-memory errors. Keep the value at a small fraction of the heap.

If you constantly see "prepared statements discarded in the last minute because cache limit reached" messages, the first step is to investigate the root cause of these messages and check whether prepared statements are used correctly - i.e. use bind markers for variable parts.

Do only change the default value, if you really have more prepared statements than fit in the cache. In most cases it is not neccessary to change this value. Constantly re-preparing statements is a performance penalty.

Default value ("auto") is 1/256th of the heap or 10MiB, whichever is greater Min unit: MiB


prepared_statements_cache_size_mb

Versions: 4.0

Default: 4.0: Auto

Maximum size of the native protocol prepared statement cache

Valid values are either "auto" (omitting the value) or a value greater 0.

Note that specifying a too large value will result in long running GCs and possbily out-of-memory errors. Keep the value at a small fraction of the heap.

If you constantly see "prepared statements discarded in the last minute because cache limit reached" messages, the first step is to investigate the root cause of these messages and check whether prepared statements are used correctly - i.e. use bind markers for variable parts.

Do only change the default value, if you really have more prepared statements than fit in the cache. In most cases it is not neccessary to change this value. Constantly re-preparing statements is a performance penalty.

Default value ("auto") is 1/256th of the heap or 10MB, whichever is greater


networking_cache_size

Versions: 4.1, 5.0

Default: 4.1: 128MiB | 5.0: 128MiB

Maximum memory to use for inter-node and client-server networking buffers.

Defaults to the smaller of 1/16 of heap or 128MB. This pool is allocated off-heap, so is in addition to the memory allocated for heap. The cache also has on-heap overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size if the default 64k chunk size is used). Memory is only allocated when needed. Min unit: MiB


networking_cache_size_in_mb

Versions: 4.0

Default: 4.0: 128

Maximum memory to use for inter-node and client-server networking buffers.

Defaults to the smaller of 1/16 of heap or 128MB. This pool is allocated off-heap, so is in addition to the memory allocated for heap. The cache also has on-heap overhead which is roughly 128 bytes per chunk (i.e. 0.2% of the reserved size if the default 64k chunk size is used). Memory is only allocated when needed.


ip_cache_max_size

Versions: 5.0

Default: 5.0: Not set

Maximum number of entries an IP to CIDR groups cache can accommodate


Snitch Configuration

endpoint_snitch

Versions: 4.0, 4.1, 5.0

Default: 4.0: SimpleSnitch | 4.1: SimpleSnitch | 5.0: SimpleSnitch

The snitch determines network topology for request routing and replica placement.

Snitch Migration

Cannot switch to incompatible snitch after data insertion. Requires adding new nodes and decommissioning old ones.

Available Snitches:

Snitch Use Case Description
SimpleSnitch Development All nodes in "datacenter1", "rack1"
GossipingPropertyFileSnitch Production Reads from cassandra-rackdc.properties
Ec2Snitch AWS single-region Region = DC, AZ = rack
Ec2MultiRegionSnitch AWS multi-region Uses public IPs
GoogleCloudSnitch GCP Region = DC, zone = rack
AzureSnitch Azure Location = DC, zone = rack

dynamic_snitch_update_interval

Versions: 4.1, 5.0

Default: 4.1: 100ms | 5.0: 100ms

controls how often to perform the more expensive part of host score calculation Min unit: ms


dynamic_snitch_update_interval_in_ms

Versions: 4.0

Default: 4.0: 100

controls how often to perform the more expensive part of host score calculation


dynamic_snitch_reset_interval

Versions: 4.1, 5.0

Default: 4.1: 600000ms | 5.0: 600000ms

controls how often to reset all host scores, allowing a bad host to possibly recover Min unit: ms


dynamic_snitch_reset_interval_in_ms

Versions: 4.0

Default: 4.0: 600000

controls how often to reset all host scores, allowing a bad host to possibly recover


dynamic_snitch_badness_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1.0 | 4.1: 1.0 | 5.0: 1.0

if set greater than zero, 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 be before the dynamic snitch will prefer other replicas over it. This is expressed as a double which represents a percentage. Thus, a value of 0.2 means Cassandra would continue to prefer the static snitch values until the pinned host was 20% worse than the fastest.


Failure Detection

phi_convict_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 8 | 4.1: 8 | 5.0: 8

phi value that must be reached for a host to be marked down. most users should never need to adjust this.


Request Timeouts

read_request_timeout

Versions: 4.1, 5.0

Default: 4.1: 5000ms | 5.0: 5000ms

How long the coordinator should wait for read operations to complete. Lowest acceptable value is 10 ms. Min unit: ms


read_request_timeout_in_ms

Versions: 4.0

Default: 4.0: 5000

How long the coordinator should wait for read operations to complete. Lowest acceptable value is 10 ms.


range_request_timeout

Versions: 4.1, 5.0

Default: 4.1: 10000ms | 5.0: 10000ms

How long the coordinator should wait for seq or index scans to complete. Lowest acceptable value is 10 ms. Min unit: ms


range_request_timeout_in_ms

Versions: 4.0

Default: 4.0: 10000

How long the coordinator should wait for seq or index scans to complete. Lowest acceptable value is 10 ms.


write_request_timeout

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

How long the coordinator should wait for writes to complete. Lowest acceptable value is 10 ms. Min unit: ms


write_request_timeout_in_ms

Versions: 4.0

Default: 4.0: 2000

How long the coordinator should wait for writes to complete. Lowest acceptable value is 10 ms.


counter_write_request_timeout

Versions: 4.1, 5.0

Default: 4.1: 5000ms | 5.0: 5000ms

How long the coordinator should wait for counter writes to complete. Lowest acceptable value is 10 ms. Min unit: ms


counter_write_request_timeout_in_ms

Versions: 4.0

Default: 4.0: 5000

How long the coordinator should wait for counter writes to complete. Lowest acceptable value is 10 ms.


cas_contention_timeout

Versions: 4.1, 5.0

Default: 4.1: 1000ms | 5.0: 1000ms

How long a coordinator should continue to retry a CAS operation that contends with other proposals for the same row. Lowest acceptable value is 10 ms. Min unit: ms


cas_contention_timeout_in_ms

Versions: 4.0

Default: 4.0: 1000

How long a coordinator should continue to retry a CAS operation that contends with other proposals for the same row. Lowest acceptable value is 10 ms.


truncate_request_timeout

Versions: 4.1, 5.0

Default: 4.1: 60000ms | 5.0: 60000ms

How long the coordinator should wait for truncates to complete (This can be much longer, because unless auto_snapshot is disabled we need to flush first so we can snapshot before removing the data.) Lowest acceptable value is 10 ms. Min unit: ms


truncate_request_timeout_in_ms

Versions: 4.0

Default: 4.0: 60000

How long the coordinator should wait for truncates to complete (This can be much longer, because unless auto_snapshot is disabled we need to flush first so we can snapshot before removing the data.) Lowest acceptable value is 10 ms.


request_timeout

Versions: 4.1, 5.0

Default: 4.1: 10000ms | 5.0: 10000ms

The default timeout for other, miscellaneous operations. Lowest acceptable value is 10 ms. Min unit: ms


request_timeout_in_ms

Versions: 4.0

Default: 4.0: 10000

The default timeout for other, miscellaneous operations. Lowest acceptable value is 10 ms.


slow_query_log_timeout

Versions: 4.1, 5.0

Default: 4.1: 500ms | 5.0: 500ms

How long before a node logs slow queries. Select queries that take longer than this timeout to execute, will generate an aggregated log message, so that slow queries can be identified. Set this value to zero to disable slow query logging. Min unit: ms


slow_query_log_timeout_in_ms

Versions: 4.0

Default: 4.0: 500

How long before a node logs slow queries. Select queries that take longer than this timeout to execute, will generate an aggregated log message, so that slow queries can be identified. Set this value to zero to disable slow query logging.


Streaming

streaming_keep_alive_period

Versions: 4.1, 5.0

Default: 4.1: 300s | 5.0: 300s

Set period for idle state control messages for earlier detection of failed streams This node will send a keep-alive message periodically on the streaming's control channel. This ensures that any eventual SocketTimeoutException will occur within 2 keep-alive cycles If the node cannot send, or timeouts sending, the keep-alive message on the netty control channel the stream session is closed. Default value is 300s (5 minutes), which means stalled streams are detected within 10 minutes Specify 0 to disable. Min unit: s


streaming_keep_alive_period_in_secs

Versions: 4.0

Default: 4.0: 300

Set keep-alive period for streaming This node will send a keep-alive message periodically with this period. If the node does not receive a keep-alive message from the peer for 2 keep-alive cycles the stream session times out and fail Default value is 300s (5 minutes), which means stalled stream times out in 10 minutes by default


stream_throughput_outbound

Versions: 4.1, 5.0

Default: 4.1: 24MiB/s | 5.0: 24MiB/s

Throttles all outbound streaming file transfers on this node to the given total throughput in Mbps. This is necessary because Cassandra does mostly sequential IO when streaming data during bootstrap or repair, which can lead to saturating the network connection and degrading rpc performance. When unset, the default is 200 Mbps or 24 MiB/s.


stream_throughput_outbound_megabits_per_sec

Versions: 4.0

Default: 4.0: 200

Throttles all outbound streaming file transfers on this node to the given total throughput in Mbps. This is necessary because Cassandra does mostly sequential IO when streaming data during bootstrap or repair, which can lead to saturating the network connection and degrading rpc performance. When unset, the default is 200 Mbps or 25 MB/s.


inter_dc_stream_throughput_outbound

Versions: 4.1, 5.0

Default: 4.1: 24MiB/s | 5.0: 24MiB/s

Throttles all streaming file transfer between the datacenters, this setting allows users to throttle inter dc stream throughput in addition to throttling all network stream traffic as configured with stream_throughput_outbound_megabits_per_sec When unset, the default is 200 Mbps or 24 MiB/s.


inter_dc_stream_throughput_outbound_megabits_per_sec

Versions: 4.0

Default: 4.0: 200

Throttles all streaming file transfer between the datacenters, this setting allows users to throttle inter dc stream throughput in addition to throttling all network stream traffic as configured with stream_throughput_outbound_megabits_per_sec When unset, the default is 200 Mbps or 25 MB/s


stream_entire_sstables

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

When enabled, permits Cassandra to zero-copy stream entire eligible SSTables between nodes, including every component. This speeds up the network transfer significantly subject to throttling specified by entire_sstable_stream_throughput_outbound, and entire_sstable_inter_dc_stream_throughput_outbound for inter-DC transfers. Enabling this will reduce the GC pressure on sending and receiving node. When unset, the default is enabled. While this feature tries to keep the disks balanced, it cannot guarantee it. This feature will be automatically disabled if internode encryption is enabled.


streaming_connections_per_host

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1 | 4.1: 1 | 5.0: 1

Limit number of connections per host for streaming Increase this when you notice that joins are CPU-bound rather that network bound (for example a few nodes with big files).


entire_sstable_stream_throughput_outbound

Versions: 4.1, 5.0

Default: 4.1: 24MiB/s | 5.0: 24MiB/s

Throttles entire SSTable outbound streaming file transfers on this node to the given total throughput in Mbps. Setting this value to 0 it disables throttling. When unset, the default is 200 Mbps or 24 MiB/s.


entire_sstable_inter_dc_stream_throughput_outbound

Versions: 4.1, 5.0

Default: 4.1: 24MiB/s | 5.0: 24MiB/s

Throttles entire SSTable file streaming between datacenters. Setting this value to 0 disables throttling for entire SSTable inter-DC file streaming. When unset, the default is 200 Mbps or 24 MiB/s.


streaming_state_expires

Versions: 4.1, 5.0

Default: 4.1: 3d | 5.0: 3d

Settings for stream stats tracking; used by system_views.streaming table How long before a stream is evicted from tracking; this impacts both historic and currently running streams.


streaming_state_size

Versions: 4.1, 5.0

Default: 4.1: 40MiB | 5.0: 40MiB

How much memory may be used for tracking before evicting session from tracking; once crossed historic and currently running streams maybe impacted.


streaming_stats_enabled

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Enable/Disable tracking of streaming stats


Security - Authentication

authenticator

Versions: 4.0, 4.1, 5.0

Default: 4.0: AllowAllAuthenticator | 4.1: AllowAllAuthenticator | 5.0: AllowAllAuthenticator

Authentication backend, implementing IAuthenticator; used to identify users Out 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 authenticate users. 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)

internode_authenticator

Versions: 4.0, 4.1, 5.0

Default: 4.0: org.apache.cassandra.auth.AllowAllInternodeAuthenticator | 4.1: org.apache.cassandra.auth.AllowAllInternodeAuthenticator | 5.0: Not set

Internode authentication backend, implementing IInternodeAuthenticator; used to allow/disallow connections from peer nodes.


Security - Authorization

authorizer

Versions: 4.0, 4.1, 5.0

Default: 4.0: AllowAllAuthorizer | 4.1: AllowAllAuthorizer | 5.0: AllowAllAuthorizer

Authorization backend, implementing IAuthorizer; used to limit access/provide permissions Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllAuthorizer, CassandraAuthorizer}.

  • AllowAllAuthorizer allows any action to any user - set it to disable authorization. - CassandraAuthorizer stores permissions in system_auth.role_permissions table. Please increase system_auth keyspace replication factor if you use this authorizer.

role_manager

Versions: 4.0, 4.1, 5.0

Default: 4.0: CassandraRoleManager | 4.1: CassandraRoleManager | 5.0: CassandraRoleManager

Part of the Authentication & Authorization backend, implementing IRoleManager; used to 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 the IRoleManager require an authenticated login, so unless the configured IAuthenticator actually implements authentication, most of this functionality will be unavailable.

  • CassandraRoleManager stores role data in the system_auth keyspace. Please increase system_auth keyspace replication factor if you use this role manager.

network_authorizer

Versions: 4.0, 4.1, 5.0

Default: 4.0: AllowAllNetworkAuthorizer | 4.1: AllowAllNetworkAuthorizer | 5.0: AllowAllNetworkAuthorizer

Network authorization backend, implementing INetworkAuthorizer; used to restrict user access to certain DCs Out of the box, Cassandra provides org.apache.cassandra.auth.{AllowAllNetworkAuthorizer, CassandraNetworkAuthorizer}.

  • AllowAllNetworkAuthorizer allows access to any DC to any user - set it to disable authorization. - CassandraNetworkAuthorizer stores permissions in system_auth.network_permissions table. Please increase system_auth keyspace replication factor if you use this authorizer.

cidr_authorizer

Versions: 5.0

Default: 5.0: (see below)

CIDR authorization backend, implementing ICIDRAuthorizer. Used to restrict user access from certain CIDRs.

cidr_authorizer:
  class_name: AllowAllCIDRAuthorizer
  parameters:
    cidr_checks_for_superusers: false
    cidr_authorizer_mode: MONITOR
    cidr_groups_cache_refresh_interval: 5
    ip_cache_max_size: 100
Sub-parameter Description Default
class_name Authorization implementation class AllowAllCIDRAuthorizer
cidr_checks_for_superusers Enable CIDR authorization for superusers (JMX calls excluded) false
cidr_authorizer_mode MONITOR (log only) or ENFORCE (reject unauthorized) MONITOR
cidr_groups_cache_refresh_interval Refresh interval for CIDR groups cache (minutes) 5
ip_cache_max_size Maximum entries in IP to CIDR groups cache 100

Implementations:

  • AllowAllCIDRAuthorizer: Allows access from any CIDR to any user. Set this to disable CIDR authorization.
  • CassandraCIDRAuthorizer: Stores user CIDR permissions in system_auth.cidr_permissions table. Increase system_auth keyspace replication factor when using this authorizer.

Modes:

  • MONITOR: CIDR checks are not enforced. Unauthorized access attempts are logged as warnings but not rejected.
  • ENFORCE: CIDR checks are enforced. Users accessing from unauthorized CIDR groups are rejected.

cidr_authorizer_mode

Versions: 5.0

Default: 5.0: MONITOR

CIDR authorizer when enabled, supports MONITOR and ENFORCE modes. Default mode is MONITOR In MONITOR mode, CIDR checks are NOT enforced. Instead, CIDR groups of users accesses are logged using nospamlogger. A warning message would be logged if a user accesses from unauthorized CIDR group (but access won't be rejected). An info message would be logged otherwise. In ENFORCE mode, CIDR checks are enforced, i.e, users accesses would be rejected if attempted from unauthorized CIDR groups.


Security - Caching

permissions_validity

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Validity period for permissions cache (fetching permissions can be an expensive operation depending on the authorizer, CassandraAuthorizer is one example). Defaults to 2000, set to 0 to disable. Will be disabled automatically for AllowAllAuthorizer. For a long-running cache using permissions_cache_active_update, consider setting to something longer such as a daily validation: 86400000ms Min unit: ms


permissions_validity_in_ms

Versions: 4.0

Default: 4.0: 2000

Validity period for permissions cache (fetching permissions can be an expensive operation depending on the authorizer, CassandraAuthorizer is one example). Defaults to 2000, set to 0 to disable. Will be disabled automatically for AllowAllAuthorizer.


permissions_update_interval

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Refresh interval for permissions cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If permissions_validity is non-zero, then this must be also. This setting is also used to inform the interval of auto-updating if using permissions_cache_active_update. Defaults to the same value as permissions_validity. For a longer-running permissions cache, consider setting to update hourly (60000) Min unit: ms


permissions_update_interval_in_ms

Versions: 4.0

Default: 4.0: 2000

Refresh interval for permissions cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If permissions_validity_in_ms is non-zero, then this must be also. Defaults to the same value as permissions_validity_in_ms.


permissions_cache_active_update

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

If true, cache contents are actively updated by a background task at the interval set by permissions_update_interval. If false, cache entries become eligible for refresh after their update interval. Upon next access, an async reload is scheduled and the old value returned until it completes.


roles_validity

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Validity period for roles cache (fetching granted roles can be an expensive operation depending on the role manager, CassandraRoleManager is one example) Granted roles are cached for authenticated sessions in AuthenticatedUser and after 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. For a long-running cache using roles_cache_active_update, consider setting to something longer such as a daily validation: 86400000 Min unit: ms


roles_validity_in_ms

Versions: 4.0

Default: 4.0: 2000

Validity period for roles cache (fetching granted roles can be an expensive operation depending on the role manager, CassandraRoleManager is one example) Granted roles are cached for authenticated sessions in AuthenticatedUser and after 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.


roles_update_interval

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Refresh interval for roles cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If roles_validity is non-zero, then this must be also. This setting is also used to inform the interval of auto-updating if using roles_cache_active_update. Defaults to the same value as roles_validity. For a long-running cache, consider setting this to 60000 (1 hour) etc. Min unit: ms


roles_update_interval_in_ms

Versions: 4.0

Default: 4.0: 2000

Refresh interval for roles cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If roles_validity_in_ms is non-zero, then this must be also. Defaults to the same value as roles_validity_in_ms.


roles_cache_active_update

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

If true, cache contents are actively updated by a background task at the interval set by roles_update_interval. If false, cache entries become eligible for refresh after their update interval. Upon next access, an async reload is scheduled and the old value returned until it completes.


credentials_validity

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Validity period for credentials cache. This cache is tightly coupled to the provided PasswordAuthenticator implementation of IAuthenticator. If another IAuthenticator implementation is configured, this cache will not be automatically used and so the following settings will have no effect. Please note, credentials are cached in their encrypted form, so while activating this cache may reduce the number of queries made to the underlying table, it may not bring a significant reduction in the latency of individual authentication attempts. Defaults to 2000, set to 0 to disable credentials caching. For a long-running cache using credentials_cache_active_update, consider setting to something longer such as a daily validation: 86400000 Min unit: ms


credentials_validity_in_ms

Versions: 4.0

Default: 4.0: 2000

Validity period for credentials cache. This cache is tightly coupled to the provided PasswordAuthenticator implementation of IAuthenticator. If another IAuthenticator implementation is configured, this cache will not be automatically used and so the following settings will have no effect. Please note, credentials are cached in their encrypted form, so while activating this cache may reduce the number of queries made to the underlying table, it may not bring a significant reduction in the latency of individual authentication attempts. Defaults to 2000, set to 0 to disable credentials caching.


credentials_update_interval

Versions: 4.1, 5.0

Default: 4.1: 2000ms | 5.0: 2000ms

Refresh interval for credentials cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If credentials_validity is non-zero, then this must be also. This setting is also used to inform the interval of auto-updating if using credentials_cache_active_update. Defaults to the same value as credentials_validity. For a longer-running permissions cache, consider setting to update hourly (60000) Min unit: ms


credentials_update_interval_in_ms

Versions: 4.0

Default: 4.0: 2000

Refresh interval for credentials cache (if enabled). After this interval, cache entries become eligible for refresh. Upon next access, an async reload is scheduled and the old value returned until it completes. If credentials_validity_in_ms is non-zero, then this must be also. Defaults to the same value as credentials_validity_in_ms.


credentials_cache_active_update

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

If true, cache contents are actively updated by a background task at the interval set by credentials_update_interval. If false (default), cache entries become eligible for refresh after their update interval. Upon next access, an async reload is scheduled and the old value returned until it completes.


auth_cache_warming_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Delays on auth resolution can lead to a thundering herd problem on reconnects; this option will enable warming of auth caches prior to node completing startup. See CASSANDRA-16958


auth_read_consistency_level

Versions: 4.1, 5.0

Default: 4.1: LOCAL_QUORUM | 5.0: LOCAL_QUORUM

configure the read and write consistency levels for modifications to auth tables


auth_write_consistency_level

Versions: 4.1, 5.0

Default: 4.1: EACH_QUORUM | 5.0: EACH_QUORUM

No description available.


Security - Encryption

server_encryption_options

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Configures server-to-server internode encryption.

Default Configuration

The default configuration is insecure. See Encryption for setup instructions.

server_encryption_options:
  internode_encryption: none
  legacy_ssl_storage_port_enabled: false
  keystore: conf/.keystore
  keystore_password: cassandra
  truststore: conf/.truststore
  truststore_password: cassandra
  require_client_auth: false
  require_endpoint_verification: false
  # For mTLS (optional)
  outbound_keystore: conf/.keystore
  outbound_keystore_password: cassandra
  # Advanced options
  protocol: TLS
  store_type: JKS
  cipher_suites: [
    TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,
    TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,
    TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA,
    TLS_RSA_WITH_AES_256_CBC_SHA
  ]
Sub-parameter Description Default
internode_encryption Encryption scope: none, dc, rack, or all none
optional Allow mixed encrypted/unencrypted connections on storage_port true if none
legacy_ssl_storage_port_enabled Open encrypted socket on ssl_storage_port (upgrade to 4.0 only) false
keystore Path to keystore file conf/.keystore
keystore_password Keystore password -
outbound_keystore Client keystore for mTLS outbound connections same as keystore
outbound_keystore_password Outbound keystore password -
truststore Path to truststore (required if require_client_auth: true) conf/.truststore
truststore_password Truststore password -
require_client_auth Require client certificates (mTLS) false
require_endpoint_verification Verify hostname in certificate matches connected host false
protocol SSL/TLS protocol TLS
store_type Keystore/truststore format JKS
cipher_suites Allowed cipher suites (list) JVM defaults
ssl_context_factory.class_name Custom SSL context factory class DefaultSslContextFactory

Internode Encryption Values:

  • none: Do not encrypt outgoing connections
  • dc: Encrypt connections to peers in other datacenters but not within datacenters
  • rack: Encrypt connections to peers in other racks but not within racks
  • all: Always use encrypted connections

Enabling Encryption (Two-Step Process):

  1. Set internode_encryption=<dc|rack|all> and optional=true. Restart all nodes.
  2. Set optional=false and optionally require_client_auth=true for mutual auth. Restart all nodes.

client_encryption_options

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Configures client-to-server encryption.

Default Configuration

The default configuration is insecure. See Encryption for setup instructions.

client_encryption_options:
  enabled: false
  keystore: conf/.keystore
  keystore_password: cassandra
  truststore: conf/.truststore
  truststore_password: cassandra
  require_client_auth: false
  require_endpoint_verification: false
  # Advanced options
  protocol: TLS
  store_type: JKS
  cipher_suites: [
    TLS_ECDHE_ECDSA_WITH_AES_256_GCM_SHA384, TLS_ECDHE_ECDSA_WITH_AES_128_GCM_SHA256,
    TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256, TLS_ECDHE_RSA_WITH_AES_128_CBC_SHA,
    TLS_ECDHE_RSA_WITH_AES_256_CBC_SHA, TLS_RSA_WITH_AES_128_GCM_SHA256, TLS_RSA_WITH_AES_128_CBC_SHA,
    TLS_RSA_WITH_AES_256_CBC_SHA
  ]
Sub-parameter Description Default
enabled Enable client-to-server encryption false
optional Allow mixed encrypted/unencrypted on native_transport_port true if disabled, false if enabled
keystore Path to keystore file conf/.keystore
keystore_password Keystore password -
truststore Path to truststore (required if require_client_auth: true) conf/.truststore
truststore_password Truststore password -
require_client_auth Require client certificates (mTLS) false
require_endpoint_verification Verify hostname in certificate false
protocol SSL/TLS protocol TLS
store_type Keystore/truststore format JKS
cipher_suites Allowed cipher suites (list) JVM defaults
ssl_context_factory.class_name Custom SSL context factory class DefaultSslContextFactory

Enabling Encryption (Two-Step Process):

  1. Set enabled=true and optional=true. Restart all nodes.
  2. Set optional=false and optionally require_client_auth=true for mutual auth. Restart all nodes.

transparent_data_encryption_options

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Configures encryption for data at rest.

transparent_data_encryption_options:
  enabled: false
  chunk_length_kb: 64
  cipher: AES/CBC/PKCS5Padding
  key_alias: testing:1
  key_provider:
    - class_name: org.apache.cassandra.security.JKSKeyProvider
      parameters:
        - keystore: conf/.keystore
          keystore_password: cassandra
Sub-parameter Description Default
enabled Enable TDE false
chunk_length_kb Encryption chunk size 64
cipher Cipher algorithm AES/CBC/PKCS5Padding
key_alias Key alias in keystore -

crypto_provider

Versions: 5.0

Default: 5.0: (see below)

Configures Java cryptography provider.

crypto_provider:
  class_name: org.apache.cassandra.security.DefaultCryptoProvider

Default installs Amazon Corretto Crypto Provider (ACCP).


ssl_context_factory

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Configure the way Cassandra creates SSL contexts. To use PEM-based key material, see org.apache.cassandra.security.PEMBasedSslContextFactory


Disk Failure Policy

disk_failure_policy

Versions: 4.0, 4.1, 5.0

Default: 4.0: stop | 4.1: stop | 5.0: stop

Policy for data disk failures:

die shut down gossip and client transports and kill the JVM for any fs errors or single-sstable errors, so the node can be replaced.

stop_paranoid shut down gossip and client transports even for single-sstable errors, kill the JVM for errors during startup.

stop shut down gossip and client transports, leaving the node effectively dead, but can still be inspected via JMX, kill the JVM for errors during startup.

best_effort stop using the failed disk and respond to requests based on remaining available sstables. This means you WILL see obsolete data at CL.ONE!

ignore ignore fatal errors and let requests fail, as in pre-1.2 Cassandra


commit_failure_policy

Versions: 4.0, 4.1, 5.0

Default: 4.0: stop | 4.1: stop | 5.0: stop

Policy for commit disk failures:

die shut down the node and kill the JVM, so the node can be replaced.

stop shut down the node, leaving the node effectively dead, but can still be inspected via JMX.

stop_commit shutdown the commit log, letting writes collect but continuing to service reads, as in pre-2.0.5 Cassandra

ignore ignore fatal errors and let the batches fail


disk_optimization_strategy

Versions: 4.0, 4.1, 5.0

Default: 4.0: ssd | 4.1: ssd | 5.0: ssd

The strategy for optimizing disk read Possible values are: ssd (for solid state disks, the default) spinning (for spinning disks)


disk_access_mode

Versions: 5.0

Default: 5.0: mmap_index_only

Policy for accessing disk:

auto Enable mmap on both data and index files on a 64-bit JVM.

standard Disable mmap entirely.

mmap Map index and data files. mmap can cause excessive paging if all actively read SSTables do not fit into RAM.

mmap_index_only Similar to mmap but maps only index files. Using this setting might also help if you observe high number of page faults or steals along with increased latencies. This setting is default.


Snapshots and Backups

incremental_backups

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Set to true to have Cassandra create a hard link to each sstable flushed or streamed locally in a backups/ subdirectory of all the keyspace data in this node. Removing these links is the operator's responsibility. The operator can also turn off incremental backups for specified table by setting table parameter incremental_backups to false, which is set to true by default. See CASSANDRA-15402


snapshot_before_compaction

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Whether or not to take a snapshot before each compaction. Be careful using this option, since Cassandra won't clean up the snapshots for you. Mostly useful if you're paranoid when there is a data format change.


auto_snapshot

Versions: 4.0, 4.1, 5.0

Default: 4.0: true | 4.1: true | 5.0: true

Whether or not a snapshot is taken of the data before keyspace truncation or dropping of column families. The STRONGLY advised default of true should be used to provide data safety. If you set this flag to false, you will lose data on truncation or drop.


auto_snapshot_ttl

Versions: 4.1, 5.0

Default: 4.1: 30d | 5.0: 30d

Adds a time-to-live (TTL) to auto snapshots generated by table truncation or drop (when enabled). After the TTL is elapsed, the snapshot is automatically cleared. By default, auto snapshots do not have TTL, uncomment the property below to enable TTL on auto snapshots. Accepted units: d (days), h (hours) or m (minutes)


Versions: 4.0, 4.1, 5.0

Default: 4.0: 0 | 4.1: 0 | 5.0: 0

The act of creating or clearing a snapshot involves creating or removing potentially tens of thousands of links, which can cause significant performance impact, especially on consumer grade SSDs. A non-zero value here can be used to throttle these links to avoid negative performance impact of taking and clearing snapshots


Repair

repair_session_space

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Limit memory usage for Merkle tree calculations during repairs of a certain table and common token range. Repair commands targetting multiple tables or virtual nodes can exceed this limit depending on concurrent_merkle_tree_requests.

The default is 1/16th of the available heap. The main tradeoff is that smaller trees have less resolution, which can lead to over-streaming data. If you see heap pressure during repairs, consider lowering this, but you cannot go below one mebibyte. If you see lots of over-streaming, consider raising this or using subrange repair.

For more details see https://issues.apache.org/jira/browse/CASSANDRA-14096.

Min unit: MiB


repair_session_space_in_mb

Versions: 4.0

Default: 4.0: Not set

Limit memory usage for Merkle tree calculations during repairs. The default is 1/16th of the available heap. The main tradeoff is that smaller trees have less resolution, which can lead to over-streaming data. If you see heap pressure during repairs, consider lowering this, but you cannot go below one megabyte. If you see lots of over-streaming, consider raising this or using subrange repair.

For more details see https://issues.apache.org/jira/browse/CASSANDRA-14096.


repaired_data_tracking_for_range_reads_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Enable tracking of repaired state of data during reads and comparison between replicas Mismatches between the repaired sets of replicas can be characterized as either confirmed or unconfirmed. In this context, unconfirmed indicates that the presence of pending repair sessions, unrepaired partition tombstones, or some other condition means that the disparity cannot be considered conclusive. Confirmed mismatches should be a trigger for investigation as they may be indicative of corruption or data loss. There are separate flags for range vs partition reads as single partition reads are only tracked when CL > 1 and a digest mismatch occurs. Currently, range queries don't use digests so if enabled for range reads, all range reads will include repaired data tracking. As this adds some overhead, operators may wish to disable it whilst still enabling it for partition reads


repaired_data_tracking_for_partition_reads_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

No description available.


report_unconfirmed_repaired_data_mismatches

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

If false, only confirmed mismatches will be reported. If true, a separate metric for unconfirmed mismatches will also be recorded. This is to avoid potential signal:noise issues are unconfirmed mismatches are less actionable than confirmed ones.


Change Data Capture (CDC)

cdc_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Enable / disable CDC functionality on a per-node basis. This modifies the logic used for write path allocation rejection (standard: never reject. cdc: reject Mutation containing a CDC-enabled table if at space limit in cdc_raw_directory).


cdc_raw_directory

Versions: 4.0, 4.1, 5.0

Default: 4.0: /var/lib/cassandra/cdc_raw | 4.1: /var/lib/cassandra/cdc_raw | 5.0: /var/lib/cassandra/cdc_raw

CommitLogSegments are moved to this directory on flush if cdc_enabled: true and the segment contains mutations for a CDC-enabled table. This should be placed on a separate spindle than the data directories. If not set, the default directory is $CASSANDRA_HOME/data/cdc_raw.


cdc_total_space

Versions: 4.1, 5.0

Default: 4.1: 4096MiB | 5.0: 4096MiB

Total space to use for change-data-capture logs on disk.

If space gets above this value, Cassandra will throw WriteTimeoutException on Mutations including tables with CDC enabled. A CDCCompactor is responsible for parsing the raw CDC logs and deleting them when parsing is completed.

The default value is the min of 4096 MiB and 1/8th of the total space of the drive where cdc_raw_directory resides. Min unit: MiB


cdc_total_space_in_mb

Versions: 4.0

Default: 4.0: 4096

Total space to use for change-data-capture logs on disk.

If space gets above this value, Cassandra will throw WriteTimeoutException on Mutations including tables with CDC enabled. A CDCCompactor is responsible for 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 space of the drive where cdc_raw_directory resides.


cdc_free_space_check_interval

Versions: 4.1, 5.0

Default: 4.1: 250ms | 5.0: 250ms

When we hit our cdc_raw limit and the CDCCompactor is either running behind or experiencing backpressure, we check at the following interval to see if any new space for cdc-tracked tables has been made available. Default to 250ms Min unit: ms


cdc_free_space_check_interval_ms

Versions: 4.0

Default: 4.0: 250

When we hit our cdc_raw limit and the CDCCompactor is either running behind or experiencing backpressure, we check at the following interval to see if any new space for cdc-tracked tables has been made available. Default to 250ms


cdc_block_writes

Versions: 5.0

Default: 5.0: true

Specify whether writes to the CDC-enabled tables should be blocked when CDC data on disk has reached to the limit. When setting to false, the writes will not be blocked and the oldest CDC data on disk will be deleted to ensure the size constraint. The default is true.


cdc_on_repair_enabled

Versions: 5.0

Default: 5.0: true

Specify whether CDC mutations are replayed through the write path on streaming, e.g. repair. When enabled, CDC data streamed to the destination node will be written into commit log first. When setting to false, the streamed CDC data is written into SSTables just the same as normal streaming. The default is true. If this is set to false, streaming will be considerably faster however it's possible that, in extreme situations (losing > quorum # nodes in a replica set), you may have data in your SSTables that never makes it to the CDC log.


User Defined Functions

user_defined_functions_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

If unset, all GC Pauses greater than gc_log_threshold will log at INFO level UDFs (user defined functions) are disabled by default. As of Cassandra 3.0 there is a sandbox in place that should prevent execution of evil code.


scripted_user_defined_functions_enabled

Versions: 4.1

Default: 4.1: false

Enables scripted UDFs (JavaScript UDFs). Java UDFs are always enabled, if user_defined_functions_enabled is true. Enable this option to be able to use UDFs with "language javascript" or any custom JSR-223 provider. This option has no effect, if user_defined_functions_enabled is false.


enable_user_defined_functions

Versions: 4.0

Default: 4.0: false

If unset, all GC Pauses greater than gc_log_threshold_in_ms will log at INFO level UDFs (user defined functions) are disabled by default. As of Cassandra 3.0 there is a sandbox in place that should prevent execution of evil code.


enable_scripted_user_defined_functions

Versions: 4.0

Default: 4.0: false

Enables scripted UDFs (JavaScript UDFs). Java UDFs are always enabled, if enable_user_defined_functions is true. Enable this option to be able to use UDFs with "language javascript" or any custom JSR-223 provider. This option has no effect, if enable_user_defined_functions is false.


Garbage Collection Logging

gc_log_threshold

Versions: 4.1, 5.0

Default: 4.1: 200ms | 5.0: 200ms

GC Pauses greater than 200 ms will be logged at INFO level This threshold can be adjusted to minimize logging if necessary Min unit: ms


gc_log_threshold_in_ms

Versions: 4.0

Default: 4.0: 200

GC Pauses greater than 200 ms will be logged at INFO level This threshold can be adjusted to minimize logging if necessary


gc_warn_threshold

Versions: 4.1, 5.0

Default: 4.1: 1000ms | 5.0: 1000ms

GC Pauses greater than gc_warn_threshold will be logged at WARN level Adjust the threshold based on your application throughput requirement. Setting to 0 will deactivate the feature. Min unit: ms


gc_warn_threshold_in_ms

Versions: 4.0

Default: 4.0: 1000

GC Pauses greater than gc_warn_threshold_in_ms will be logged at WARN level Adjust the threshold based on your application throughput requirement. Setting to 0 will deactivate the feature.


Tombstones

tombstone_warn_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 1000 | 4.1: 1000 | 5.0: 1000

When executing a scan, within or across a partition, we need to keep the tombstones seen in memory so we can return them to the coordinator, which will use them to make sure other replicas also know about the deleted rows. With workloads that generate a lot of tombstones, this can cause performance problems 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 to scan more tombstones anyway. These thresholds may also be adjusted at runtime using the StorageService mbean.


tombstone_failure_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 100000 | 4.1: 100000 | 5.0: 100000

No description available.


Batch Operations

batch_size_warn_threshold

Versions: 4.1, 5.0

Default: 4.1: 5KiB | 5.0: 5KiB

Log WARN on any multiple-partition batch size exceeding this value. 5KiB per batch by default. Caution should be taken on increasing the size of this threshold as it can lead to node instability. Min unit: KiB


batch_size_warn_threshold_in_kb

Versions: 4.0

Default: 4.0: 5

Log 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.


batch_size_fail_threshold

Versions: 4.1, 5.0

Default: 4.1: 50KiB | 5.0: 50KiB

Fail any multiple-partition batch exceeding this value. 50KiB (10x warn threshold) by default. Min unit: KiB


batch_size_fail_threshold_in_kb

Versions: 4.0

Default: 4.0: 50

Fail any multiple-partition batch exceeding this value. 50kb (10x warn threshold) by default.


unlogged_batch_across_partitions_warn_threshold

Versions: 4.0, 4.1, 5.0

Default: 4.0: 10 | 4.1: 10 | 5.0: 10

Log WARN on any batches not of type LOGGED than span across more partitions than this limit


batchlog_replay_throttle

Versions: 4.1, 5.0

Default: 4.1: 1024KiB | 5.0: 1024KiB

Maximum throttle in KiBs per second, total. This will be reduced proportionally to the number of nodes in the cluster. Min unit: KiB


batchlog_replay_throttle_in_kb

Versions: 4.0

Default: 4.0: 1024

Maximum throttle in KBs per second, total. This will be reduced proportionally to the number of nodes in the cluster.


batchlog_endpoint_strategy

Versions: 5.0

Default: 5.0: random_remote

Strategy to choose the batchlog storage endpoints.

Available options:

  • random_remote Default, purely random, prevents the local rack, if possible.

  • prefer_local Similar to random_remote. Random, except that one of the replications will go to the local rack, which mean it offers lower availability guarantee than random_remote or dynamic_remote.

  • dynamic_remote Using DynamicEndpointSnitch to select batchlog storage endpoints, prevents the local rack, if possible. This strategy offers the same availability guarantees as random_remote but selects the fastest endpoints according to the DynamicEndpointSnitch. (DynamicEndpointSnitch currently only tracks reads and not writes - i.e. write-only (or mostly-write) workloads might not benefit from this strategy.) Note: this strategy will fall back to random_remote, if dynamic_snitch is not enabled.

  • dynamic Mostly the same as dynamic_remote, except that local rack is not excluded, which mean it offers lower availability guarantee than random_remote or dynamic_remote. Note: this strategy will fall back to random_remote, if dynamic_snitch is not enabled.


Query Processing

column_index_size

Versions: 4.1, 5.0

Default: 4.1: 64KiB | 5.0: 4KiB

Granularity of the collation index of rows within a partition. Applies to both BIG and BTI SSTable formats. In both formats, a smaller granularity results in faster lookup of rows within a partition, but a bigger index file size. Using smaller granularities with the BIG format is not recommended because bigger collation indexes cannot be cached efficiently or at all if they become sufficiently large. Further, if large rows, or a very large number of rows per partition are present, it is recommended to increase the index granularity or switch to the BTI SSTable format.

Leave undefined to use a default suitable for the SSTable format in use (64 KiB for BIG, 16KiB for BTI). Min unit: KiB


column_index_size_in_kb

Versions: 4.0

Default: 4.0: 64

Granularity of the collation index of rows within a partition. Increase if your rows are large, or if you have a very large number of rows per partition. The competing goals are these:

  • a smaller granularity means more index entries are generated and looking up rows withing the partition by collation column is faster - but, Cassandra will keep the collation index in memory for hot rows (as part of the key cache), so a larger granularity means you can cache more hot rows

column_index_cache_size

Versions: 4.1, 5.0

Default: 4.1: 2KiB | 5.0: 2KiB

Per sstable indexed key cache entries (the collation index in memory mentioned above) exceeding this size will not be held on heap. This means that only partition information is held on heap and the index entries are read from disk.

Note that this size refers to the size of the serialized index information and not the size of the partition.

This is only relevant to SSTable formats that use key cache, e.g. BIG. Min unit: KiB


column_index_cache_size_in_kb

Versions: 4.0

Default: 4.0: 2

Per sstable indexed key cache entries (the collation index in memory mentioned above) exceeding this size will not be held on heap. This means that only partition information is held on heap and the index entries are read from disk.

Note that this size refers to the size of the serialized index information and not the size of the partition.


enable_materialized_views

Versions: 4.0

Default: 4.0: false

Enables materialized view creation on this node. Materialized views are considered experimental and are not recommended for production use.


enable_sasi_indexes

Versions: 4.0

Default: 4.0: false

Enables SASI index creation on this node. SASI indexes are considered experimental and are not recommended for production use.


sasi_indexes_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Enables SASI index creation on this node. SASI indexes are considered experimental and are not recommended for production use.


enable_transient_replication

Versions: 4.0

Default: 4.0: false

Enables creation of transiently replicated keyspaces on this node. Transient replication is experimental and is not recommended for production use.


transient_replication_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Enables creation of transiently replicated keyspaces on this node. Transient replication is experimental and is not recommended for production use.


enable_drop_compact_storage

Versions: 4.0

Default: 4.0: false

Enables the used of 'ALTER ... DROP COMPACT STORAGE' statements on this node. 'ALTER ... DROP COMPACT STORAGE' is considered experimental and is not recommended for production use.


drop_compact_storage_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Enables the used of 'ALTER ... DROP COMPACT STORAGE' statements on this node. 'ALTER ... DROP COMPACT STORAGE' is considered experimental and is not recommended for production use.


ideal_consistency_level

Versions: 4.0, 4.1, 5.0

Default: 4.0: EACH_QUORUM | 4.1: EACH_QUORUM | 5.0: EACH_QUORUM

Track a metric per keyspace indicating whether replication achieved the ideal consistency level for writes without timing out. This is different from the consistency level requested by each write which may be lower in order to facilitate availability.


traverse_auth_from_root

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Depending on the auth strategy of the cluster, it can be beneficial to iterate from root to table (root -> ks -> table) instead of table to root (table -> ks -> root). As the auth entries are whitelisting, once a permission is found you know it to be valid. We default to false as the legacy behavior is to query at the table level then move back up to the root. See CASSANDRA-17016 for details.


use_statements_enabled

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Whether or not USE is allowed. This is enabled by default to avoid failure on upgrade.


read_before_write_list_operations_enabled

Versions: 4.1, 5.0

Default: 4.1: true | 5.0: true

Guardrail to allow/disallow list operations that require read before write, i.e. setting list element by index and removing list elements by either index or value. Defaults to true.


replica_filtering_protection

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Protects against out-of-date replicas causing memory issues.

replica_filtering_protection:
  cached_rows_warn_threshold: 2000
  cached_rows_fail_threshold: 32000
Sub-parameter Description Default
cached_rows_warn_threshold Warning threshold 2000
cached_rows_fail_threshold Failure threshold 32000

materialized_views_enabled

Versions: 4.1, 5.0

Default: 4.1: false | 5.0: false

Enables materialized view creation on this node. Materialized views are considered experimental and are not recommended for production use.


Indexing

default_secondary_index

Versions: 5.0

Default: 5.0: legacy_local_table

The default secondary index implementation when CREATE INDEX does not specify one via USING. ex. "legacy_local_table" - (default) legacy secondary index, implemented as a hidden table ex. "sai" - "storage-attched" index, implemented via optimized SSTable/Memtable-attached indexes


default_secondary_index_enabled

Versions: 5.0

Default: 5.0: true

Whether a default secondary index implementation is allowed. If this is "false", CREATE INDEX must specify an index implementation via USING.


sai_options

Versions: 5.0

Default: 5.0: Not set

Configures Storage Attached Indexing (SAI).

sai_options:
  segment_write_buffer_size: 1024MiB
Sub-parameter Description Default
segment_write_buffer_size Memory for SAI segments 1024MiB
prioritize_over_legacy_index Prefer SAI over legacy false

Audit Logging

audit_logging_options

Versions: 4.0, 4.1, 5.0

Default: 4.0: (see below) | 4.1: (see below) | 5.0: (see below)

Configures audit logging for tracking database operations.

audit_logging_options:
  enabled: false
  logger:
    - class_name: BinAuditLogger
  audit_logs_dir: /var/log/cassandra/audit
  roll_cycle: HOURLY
  block: true
  max_queue_weight: 268435456
  max_log_size: 17179869184
  max_archive_retries: 10
Sub-parameter Description Default
enabled Enable audit logging false
logger.class_name Logger implementation: BinAuditLogger or FileAuditLogger BinAuditLogger
audit_logs_dir Directory for audit log files -
included_keyspaces Keyspaces to include (empty = all) -
excluded_keyspaces Keyspaces to exclude system, system_schema, system_virtual_schema
included_categories Categories to include (empty = all) -
excluded_categories Categories to exclude -
included_users Users to include (empty = all) -
excluded_users Users to exclude -
roll_cycle Log rotation: MINUTELY, HOURLY, DAILY HOURLY
block Block writes when queue is full true
max_queue_weight Maximum queue size in bytes 268435456 (256 MiB)
max_log_size Maximum total log size in bytes 17179869184 (16 GiB)
archive_command Script to run on log rotation (%path = file path) -
max_archive_retries Maximum archive command retries 10

Archiving Behavior

If archive_command is empty, Cassandra uses a built-in DeletingArchiver that deletes oldest files when max_log_size is reached. If set, the script is responsible for cleanup.


full_query_logging_options

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

Configures full query logging. These defaults can be overridden via nodetool enablefullquerylog.

full_query_logging_options:
  log_dir: /var/log/cassandra/fql
  roll_cycle: HOURLY
  block: true
  max_queue_weight: 268435456
  max_log_size: 17179869184
  max_archive_retries: 10
Sub-parameter Description Default
log_dir Directory for query log files -
roll_cycle Log rotation: MINUTELY, HOURLY, DAILY HOURLY
block Block writes when queue is full true
max_queue_weight Maximum queue size in bytes 268435456 (256 MiB)
max_log_size Maximum total log size in bytes 17179869184 (16 GiB)
archive_command Script to run on log rotation (%path = file path) -
allow_nodetool_archive_command Allow nodetool to set archive_command (security risk) false
max_archive_retries Maximum archive command retries 10

Security

Enabling allow_nodetool_archive_command allows anyone with JMX/nodetool access to run local shell commands as the Cassandra user.


Tracing

trace_type_query_ttl

Versions: 4.1, 5.0

Default: 4.1: 1d | 5.0: 1d

TTL for different trace types used during logging of the repair process. Min unit: s


trace_type_repair_ttl

Versions: 4.1, 5.0

Default: 4.1: 7d | 5.0: 7d

Min unit: s


tracetype_query_ttl

Versions: 4.0

Default: 4.0: 86400

TTL for different trace types used during logging of the repair process.


tracetype_repair_ttl

Versions: 4.0

Default: 4.0: 604800

No description available.


Diagnostic Events

diagnostic_events_enabled

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Diagnostic Events # If enabled, diagnostic events can be helpful for troubleshooting operational issues. Emitted events contain details on internal state and temporal relationships across events, accessible by clients via JMX.


Index Summary

index_summary_capacity

Versions: 4.1, 5.0

Default: 4.1: Auto | 5.0: Auto

A fixed memory pool size in MB for for SSTable index summaries. If left empty, this will default to 5% of the heap size. If the memory usage of all index summaries exceeds this limit, SSTables with low read rates will shrink their index summaries in order to meet this limit. However, this is a best-effort process. In extreme conditions Cassandra may need to use more than this amount of memory. Only relevant to formats that use an index summary, e.g. BIG. Min unit: KiB


index_summary_capacity_in_mb

Versions: 4.0

Default: 4.0: Auto

A fixed memory pool size in MB for for SSTable index summaries. If left empty, this will default to 5% of the heap size. If the memory usage of all index summaries exceeds this limit, SSTables with low read rates will shrink their index summaries in order to meet this limit. However, this is a best-effort process. In extreme conditions Cassandra may need to use more than this amount of memory.


index_summary_resize_interval

Versions: 4.1, 5.0

Default: 4.1: 60m | 5.0: 60m

How frequently index summaries should be resampled. This is done periodically to redistribute memory from the fixed-size pool to sstables proportional their recent read rates. Setting to null value will disable this process, leaving existing index summaries at their current sampling level. Only relevant to formats that use an index summary, e.g. BIG. Min unit: m


index_summary_resize_interval_in_minutes

Versions: 4.0

Default: 4.0: 60

How frequently index summaries should be resampled. This is done periodically to redistribute memory from the fixed-size pool to sstables proportional their recent read rates. Setting to -1 will disable this process, leaving existing index summaries at their current sampling level.


Trickle Fsync

trickle_fsync

Versions: 4.0, 4.1, 5.0

Default: 4.0: false | 4.1: false | 5.0: false

Whether to, when doing sequential writing, fsync() at intervals in order to force the operating system to flush the dirty buffers. Enable this to avoid sudden dirty buffer flushing from impacting read latencies. Almost always a good idea on SSDs; not necessarily on platters.


trickle_fsync_interval

Versions: 4.1, 5.0

Default: 4.1: 10240KiB | 5.0: 10240KiB

Min unit: KiB


trickle_fsync_interval_in_kb

Versions: 4.0

Default: 4.0: 10240

No description available.


Windows

windows_timer_interval

Versions: 4.0

Default: 4.0: 1

The default Windows kernel timer and scheduling resolution is 15.6ms for power conservation. Lowering this value on Windows can provide much tighter latency and better throughput, however some virtualized environments may see a negative performance impact from changing this setting below their system default. The sysinternals 'clockres' tool can confirm your system's default setting.


Startup Checks

startup_checks

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Controls startup validation checks.

startup_checks:
  enabled: true
  check_dc: true
  check_rack: true
Sub-parameter Description Default
enabled Enable checks true
check_dc Verify datacenter true
check_rack Verify rack true

check_filesystem_ownership

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Verifies correct ownership of attached locations on disk at startup. See CASSANDRA-16879 for more details.


check_data_resurrection

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

Enable this property to fail startup if the node is down for longer than gc_grace_seconds, to potentially prevent data resurrection on tables with deletes. By default, this will run against all keyspaces and tables except the ones specified on excluded_keyspaces and excluded_tables.


Client Error Reporting

client_error_reporting_exclusions

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

When the client triggers a protocol exception or unknown issue (Cassandra bug) we increment a client metric showing this; this logic will exclude specific subnets from updating these metrics


Storage Compatibility

storage_compatibility_mode

Versions: 5.0

Default: 5.0: CASSANDRA_4

This property indicates with what Cassandra major version the storage format will be compatible with.

The chosen storage compatibility mode will determine the versions of the written sstables, commitlogs, hints, etc. For example, if we're going to remain compatible with Cassandra 4.x, the value of this property should be 4, which will make us use sstables in the latest N version of the BIG format.

This will also determine if certain features that depend on newer formats are available. For example, extended TTL (up to 2106) depends on the sstable, commit-log, hints, and messaging versions introduced by Cassandra 5.0, so that feature won't be available if this property is set to CASSANDRA_4. See the upgrade guide for more details.

Possible values are:

  • CASSANDRA_4: Stays compatible with the 4.x line in features, formats and component versions.
  • UPGRADING: The cluster monitors the version of each node during this interim stage. This has a cost but ensures all new features, formats, versions, etc. are enabled safely.
  • NONE: Start with all the new features and formats enabled.

A typical upgrade would be:

  1. Do a rolling upgrade, starting all nodes in CASSANDRA_X compatibility mode.
  2. Once the new binary is rendered stable, do a rolling restart with the UPGRADING mode. The cluster will keep new features disabled until all nodes are started in the UPGRADING mode; when that happens, new features controlled by the storage compatibility mode are enabled.
  3. Do a rolling restart with all nodes starting with the NONE mode. This eliminates the cost of checking node versions and ensures stability. If Cassandra was started at the previous version by accident, a node with disabled compatibility mode would no longer toggle behaviors as when it was running in the UPGRADING mode.

Dynamic Data Masking

dynamic_data_masking_enabled

Versions: 5.0

Default: 5.0: false

If enabled, dynamic data masking allows to attach CQL masking functions to the columns of a table. Users without the UNMASK permission will see an obscured version of the values of the columns with an attached mask. If dynamic data masking is disabled it won't be allowed to create new column masks, although it will still be possible to drop any previously existing masks. Also, any existing mask will be ignored at query time, so all users will see the clear values of the masked columns. Defaults to false to disable dynamic data masking.


Corrupted Tombstones

corrupted_tombstone_strategy

Versions: 4.0, 4.1, 5.0

Default: 4.0: disabled | 4.1: disabled | 5.0: disabled

validate tombstones on reads and compaction can be either "disabled", "warn" or "exception"


DNS

resolve_multiple_ip_addresses_per_dns_record

Versions: 5.0

Default: 5.0: Not set

If set to "true", SimpleSeedProvider will return all IP addresses for a DNS name, based on the configured name service on the system. Defaults to "false".


Archive

archive_command

Versions: 4.0, 4.1, 5.0

Default: 4.0: Not set | 4.1: Not set | 5.0: Not set

If archive_command is empty or unset, Cassandra uses a built-in DeletingArchiver that deletes the oldest files if max_log_size is reached. If archive_command is set, Cassandra does not use DeletingArchiver, so it is the responsibility of the script to make any required cleanup. Example: "/path/to/script.sh %path" where %path is replaced with the file being rolled.


allow_nodetool_archive_command

Versions: 4.1, 5.0

Default: 4.1: Not set | 5.0: Not set

note that enabling this allows anyone with JMX/nodetool access to run local shell commands as the user running cassandra


Heap Dump

dump_heap_on_uncaught_exception

Versions: 5.0

Default: 5.0: true

Enable / disable automatic dump of heap on first uncaught exception If not set, the default value is false


heap_dump_path

Versions: 5.0

Default: 5.0: /var/lib/cassandra/heapdump

Directory where Cassandra should store results of a One-Shot troubleshooting heapdump for uncaught exceptions. Note: this value can be overridden by the -XX:HeapDumpPath JVM env param with a relative local path for testing if so desired. If not set, the default directory is $CASSANDRA_HOME/heapdump


Guardrails

Guardrails protect the cluster by warning or failing operations that exceed configured thresholds. Most guardrails were introduced in Cassandra 4.1 with additional ones added in 5.0.

Guardrail Types

  • warn_threshold: Logs a warning but allows the operation
  • fail_threshold: Rejects the operation with an error
  • _enabled: Toggles the feature on/off
  • _warned: Warns on specific values
  • _disallowed: Rejects specific values

Schema Guardrails

Parameter Description Versions
keyspaces_warn_threshold Warn when creating more keyspaces than threshold 4.1, 5.0
keyspaces_fail_threshold Fail when creating more keyspaces than threshold 4.1, 5.0
tables_warn_threshold Warn when creating more tables than threshold 4.1, 5.0
tables_fail_threshold Fail when creating more tables than threshold 4.1, 5.0
columns_per_table_warn_threshold Warn when table has more columns than threshold 4.1, 5.0
columns_per_table_fail_threshold Fail when table has more columns than threshold 4.1, 5.0
fields_per_udt_warn_threshold Warn when UDT has more fields than threshold 4.1, 5.0
fields_per_udt_fail_threshold Fail when UDT has more fields than threshold 4.1, 5.0

Index Guardrails

Parameter Description Versions
secondary_indexes_enabled Enable/disable secondary index creation 4.1, 5.0
secondary_indexes_per_table_warn_threshold Warn when table has more secondary indexes than threshold 4.1, 5.0
secondary_indexes_per_table_fail_threshold Fail when table has more secondary indexes than threshold 4.1, 5.0
materialized_views_per_table_warn_threshold Warn when table has more MVs than threshold 4.1, 5.0
materialized_views_per_table_fail_threshold Fail when table has more MVs than threshold 4.1, 5.0
non_partition_restricted_index_query_enabled Enable/disable non-partition-restricted index queries 5.0

SAI Guardrails (5.0+)

Parameter Description Versions
sai_string_term_size_warn_threshold Warn on large SAI string terms 5.0
sai_string_term_size_fail_threshold Fail on large SAI string terms 5.0
sai_frozen_term_size_warn_threshold Warn on large SAI frozen terms 5.0
sai_frozen_term_size_fail_threshold Fail on large SAI frozen terms 5.0
sai_vector_term_size_warn_threshold Warn on large SAI vector terms 5.0
sai_vector_term_size_fail_threshold Fail on large SAI vector terms 5.0
sai_sstable_indexes_per_query_warn_threshold Warn when query touches many SAI indexes 5.0
sai_sstable_indexes_per_query_fail_threshold Fail when query touches many SAI indexes 5.0

Data Size Guardrails

Parameter Description Versions
partition_size_warn_threshold Warn on partitions larger than threshold 5.0
partition_size_fail_threshold Fail on partitions larger than threshold 5.0
partition_tombstones_warn_threshold Warn on partitions with too many tombstones 5.0
partition_tombstones_fail_threshold Fail on partitions with too many tombstones 5.0
collection_size_warn_threshold Warn on collections larger than threshold 4.1, 5.0
collection_size_fail_threshold Fail on collections larger than threshold 4.1, 5.0
items_per_collection_warn_threshold Warn on collections with too many items 4.1, 5.0
items_per_collection_fail_threshold Fail on collections with too many items 4.1, 5.0
column_value_size_warn_threshold Warn on column values larger than threshold 5.0
column_value_size_fail_threshold Fail on column values larger than threshold 5.0

Query Guardrails

Parameter Description Versions
page_size_warn_threshold Warn on queries with page size larger than threshold 4.1, 5.0
page_size_fail_threshold Fail on queries with page size larger than threshold 4.1, 5.0
partition_keys_in_select_warn_threshold Warn when IN restriction has too many partition keys 4.1, 5.0
partition_keys_in_select_fail_threshold Fail when IN restriction has too many partition keys 4.1, 5.0
in_select_cartesian_product_warn_threshold Warn when IN creates large cartesian product 4.1, 5.0
in_select_cartesian_product_fail_threshold Fail when IN creates large cartesian product 4.1, 5.0
allow_filtering_enabled Enable/disable ALLOW FILTERING queries 4.1, 5.0
group_by_enabled Enable/disable GROUP BY functionality 4.1, 5.0

Read Size Guardrails

Parameter Description Versions
local_read_size_warn_threshold Warn on reads larger than threshold at local node 4.1, 5.0
local_read_size_fail_threshold Fail on reads larger than threshold at local node 4.1, 5.0
coordinator_read_size_warn_threshold Warn on reads larger than threshold at coordinator 4.1, 5.0
coordinator_read_size_fail_threshold Fail on reads larger than threshold at coordinator 4.1, 5.0
row_index_read_size_warn_threshold Warn on row index reads larger than threshold 4.1, 5.0
row_index_read_size_fail_threshold Fail on row index reads larger than threshold 4.1, 5.0
read_thresholds_enabled Enable/disable read size thresholds 4.1, 5.0

Consistency Level Guardrails

Parameter Description Versions
read_consistency_levels_warned Warn on these read consistency levels 4.1, 5.0
read_consistency_levels_disallowed Disallow these read consistency levels 4.1, 5.0
write_consistency_levels_warned Warn on these write consistency levels 4.1, 5.0
write_consistency_levels_disallowed Disallow these write consistency levels 4.1, 5.0

Replication Guardrails

Parameter Description Versions
minimum_replication_factor_warn_threshold Warn when RF below threshold 4.1, 5.0
minimum_replication_factor_fail_threshold Fail when RF below threshold 4.1, 5.0
maximum_replication_factor_warn_threshold Warn when RF above threshold 5.0
maximum_replication_factor_fail_threshold Fail when RF above threshold 5.0
simplestrategy_enabled Enable/disable SimpleStrategy 5.0

Timestamp Guardrails

Parameter Description Versions
user_timestamps_enabled Enable/disable user-provided timestamps 4.1, 5.0
minimum_timestamp_warn_threshold Warn on timestamps below threshold 5.0
minimum_timestamp_fail_threshold Fail on timestamps below threshold 5.0
maximum_timestamp_warn_threshold Warn on timestamps above threshold 5.0
maximum_timestamp_fail_threshold Fail on timestamps above threshold 5.0

DDL Guardrails

Parameter Description Versions
drop_truncate_table_enabled Enable/disable DROP/TRUNCATE table 4.1, 5.0
drop_keyspace_enabled Enable/disable DROP KEYSPACE 5.0
alter_table_enabled Enable/disable ALTER TABLE 5.0
uncompressed_tables_enabled Enable/disable uncompressed tables 4.1, 5.0
table_properties_warned Warn on these table properties 4.1, 5.0
table_properties_disallowed Disallow these table properties 4.1, 5.0

Partition Denylist

Parameter Description Versions
partition_denylist_enabled Enable partition denylist feature 4.1, 5.0
denylist_reads_enabled Enable read operations on denylisted partitions 4.1, 5.0
denylist_writes_enabled Enable write operations on denylisted partitions 4.1, 5.0
denylist_range_reads_enabled Enable range reads on denylisted partitions 4.1, 5.0

TWCS Guardrails

Parameter Description Versions
zero_ttl_on_twcs_enabled Enable/disable zero TTL on TWCS tables 5.0
zero_ttl_on_twcs_warned Warn on zero TTL on TWCS tables 5.0

Vector Guardrails (5.0+)

Parameter Description Versions
vector_dimensions_warn_threshold Warn on vectors with too many dimensions 5.0
vector_dimensions_fail_threshold Fail on vectors with too many dimensions 5.0

Disk Usage Guardrails

Parameter Description Versions
data_disk_usage_percentage_warn_threshold Warn when disk usage exceeds percentage 4.1, 5.0
data_disk_usage_percentage_fail_threshold Fail when disk usage exceeds percentage 4.1, 5.0