Index of wsrep_provider options

The following variables can be set and checked in the wsrep_provider_options variable. The value of the variable can be changed in the MySQL configuration file, my.cnf, or by setting the variable value in the MySQL client.

To change the value in my.cnf, the following syntax should be used:

wsrep_provider_options="variable1=value1;[variable2=value2]"

For example to set the size of the Galera buffer storage to 512 MB, specify the following in my.cnf:

wsrep_provider_options="gcache.size=512M"

Dynamic variables can be changed from the MySQL client using the SET GLOBAL command. For example, to change the value of the pc.ignore_sb, use the following command:

mysql> SET GLOBAL wsrep_provider_options="pc.ignore_sb=true";

Index

base_dir

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	value of datadir

This variable specifies the data directory.

base_host

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	value of wsrep_node_address

This variable sets the value of the node’s base IP. This is an IP address on which Galera listens for connections from other nodes. Setting this value incorrectly would stop the node from communicating with other nodes.

base_port

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	4567

This variable sets the port on which Galera listens for connections from other nodes. Setting this value incorrectly would stop the node from communicating with other nodes.

cert.log_conflicts

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	no

This variable is used to specify if the details of the certification failures should be logged.

debug

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	no

When this variable is set to yes, it will enable debugging.

evs.auto_evict

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	0

Number of entries allowed on delayed list until auto eviction takes place. Setting value to 0 disables auto eviction protocol on the node, though node response times will still be monitored. EVS protocol version (evs.version) 1 is required to enable auto eviction.

evs.causal_keepalive_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	value of evs.keepalive_period

This variable is used for development purposes and shouldn’t be used by regular users.

evs.debug_log_mask

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	0x1

This variable is used for EVS (Extended Virtual Synchrony) debugging. It can be used only when wsrep_debug is set to ON.

evs.delay_margin

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	PT1S

Time period that a node can delay its response from expected until it is added to delayed list. The value must be higher than the highest RTT between nodes.

evs.delayed_keep_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	PT30S

Time period that node is required to remain responsive until one entry is removed from delayed list.

evs.evict

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes

Manual eviction can be triggered by setting the evs.evict to a certain node value. Setting the evs.evict to an empty string will clear the evict list on the node where it was set.

evs.inactive_check_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT0.5S

This variable defines how often to check for peer inactivity.

evs.inactive_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT15S

This variable defines the inactivity limit, once this limit is reached the node will be considered dead.

evs.info_log_mask

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable is used for controlling the extra EVS info logging.

evs.install_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	PT7.5S

This variable defines the timeout on waiting for install message acknowledgments.

evs.join_retrans_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT1S

This variable defines how often to retransmit EVS join messages when forming cluster membership.

evs.keepalive_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT1S

This variable defines how often to emit keepalive beacons (in the absence of any other traffic).

evs.max_install_timeouts

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	1

This variable defines how many membership install rounds to try before giving up (total rounds will be evs.max_install_timeouts + 2).

evs.send_window

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	10

This variable defines the maximum number of data packets in replication at a time. For WAN setups, the variable can be set to a considerably higher value than default (for example,512). The value must not be less than evs.user_send_window.

evs.stats_report_period

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT1M

This variable defines the control period of EVS statistics reporting.

evs.suspect_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT5S

This variable defines the inactivity period after which the node is “suspected” to be dead. If all remaining nodes agree on that, the node will be dropped out of cluster even before evs.inactive_timeout is reached.

evs.use_aggregate

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	true

When this variable is enabled, smaller packets will be aggregated into one.

evs.user_send_window

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	4

This variable defines the maximum number of data packets in replication at a time. For WAN setups, the variable can be set to a considerably higher value than default (for example, 512).

evs.version

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable defines the EVS protocol version. Auto eviction is enabled when this variable is set to 1. Default 0 is set for backwards compatibility.

evs.view_forget_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	P1D

This variable defines the timeout after which past views will be dropped from history.

gcache.dir

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	datadir

This variable can be used to define the location of the galera.cache file.

gcache.freeze_purge_at_seqno

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Local, Global
Dynamic:	Yes
Default Value:	0

This variable controls the purging of the gcache and enables retaining more data in it. This variable makes it possible to use IST (Incremental State Transfer) when the node rejoins instead of SST (State Snapshot Transfer).

Set this variable on an existing node of the cluster (that will continue to be part of the cluster and can act as a potential donor node). This node continues to retain the write-sets and allows restarting the node to rejoin by using IST.

See also

Percona Database Performance Blog:

• All You Need to Know About GCache (Galera-Cache)

• Want IST Not SST for Node Rejoins? We Have a Solution!

The gcache.freeze_purge_at_seqno variable takes three values:

-1 (default)

No freezing of gcache, the purge operates as normal.

A valid seqno in gcache

The freeze purge of write-sets may not be smaller than the selected seqno. The best way to select an optimal value is to use the value of the variable :variable:wsrep_last_applied from the node that you plan to shut down.

now The freeze purge of write-sets is no less than the smallest seqno currently in gcache. Using this value results in freezing the gcache-purge instantly. Use this value if selecting a valid seqno in gcache is difficult.

gcache.keep_pages_count

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Local, Global
Dynamic:	Yes
Default Value:	0

This variable is used to limit the number of overflow pages rather than the total memory occupied by all overflow pages. Whenever gcache.keep_pages_count is set to a non-zero value, excess overflow pages will be deleted (starting from the oldest to the newest).

Whenever either the gcache.keep_pages_count or the gcache.keep_pages_size variable is updated at runtime to a non-zero value, cleanup is called on excess overflow pages to delete them.

gcache.keep_pages_size

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Local, Global
Dynamic:	No
Default Value:	0

This variable is used to limit the total size of overflow pages rather than the count of all overflow pages. Whenever gcache.keep_pages_size is set to a non-zero value, excess overflow pages will be deleted (starting from the oldest to the newest) until the total size is below the specified value.

Whenever either the gcache.keep_pages_count or the gcache.keep_pages_size variable is updated at runtime to a non-zero value, cleanup is called on excess overflow pages to delete them.

gcache.mem_size

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable has been deprecated in 5.6.22-25.8 and shouldn’t be used as it could cause a node to crash.

This variable was used to define how much RAM is available for the system.

gcache.name

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	/var/lib/mysql/galera.cache

This variable can be used to specify the name of the Galera cache file.

gcache.page_size

Option	Description
Command Line:	No
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	128M

Size of the page files in page storage. The limit on overall page storage is the size of the disk. Pages are prefixed by gcache.page.

See also

• Galera Documentation: gcache.page_size

• [Percona Database Performance Blog: All You Need to Know About GCache (https://www.percona.com/blog/2016/11/16/all-you-need-to-know-about-gcache-galera-cache/)

gcache.recover

Option	Description
Command Line:	No
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	No

Attempts to recover a node’s gcache file to a usable state on startup. If the node can successfully recover the gcache file, the node can provide IST to the remaining nodes. This ability can reduce the time needed to bring up the cluster.

An example of setting the value to yes in the configuration file:

wsrep_provider_options="gcache.recover=yes"

gcache.size

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	128M

Size of the transaction cache for Galera replication. This defines the size of the galera.cache file which is used as source for IST. The bigger the value of this variable, the better are chances that the re-joining node will get IST instead of SST.

gcomm.thread_prio

Raises the gcomm thread priority to a higher level. Use this variable when the gcomm thread does not receive enough CPU time due to other competing threads. For example, if the gcomm threads are not frequently run, a node may drop from the cluster because of the timeout.

The format for this variable is: <policy>:<priority>. The policy value supports the following options: other, fifo, and rr. The priority value is an integer.

Note

Setting a priority value of 99 is not recommended. This value blocks system threads.

An example of the variable:

wsrep_provider_options="gcomm.thread_prio=fifo:3"

The description of the policy parameter follows:

Option	Description
other	This policy is the default Linux time-sharing scheduling. Threads run until one of the following events occur: * Thread exit I/O request blocks the thread Higher priority thread preempts the thread
fifo	The policy uses a First-in First-out (FIFO) scheduling. These threads always immediately preempt any currently running other, batch or idle threads. The threads are run in a FIFO manner until completion, unless a higher priority thread preempts or blocks them. This policy does not use time slicing.
rr	The threads use round-robin scheduling. This thread always preempts a currently running other, batch or idle thread. The scheduler runs threads with the same priority for a fixed time in a round-robin style. When this time period is exceeded, the scheduler stops the thread and moves it to the end of the list, and runs another round-robin thread with the same priority.

See also

For information, see the Galera Cluster documentation

gcs.fc_debug

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable specifies after how many writesets the debug statistics about SST flow control will be posted.

gcs.fc_factor

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	1

This variable is used for replication flow control. Replication is resumed when the replica queue drops below gcs.fc_factor * gcs.fc_limit.

gcs.fc_limit

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	100

The dafaul velue has been changed from 16 to 100 in 5.7.17-29.20.

This variable is used for replication flow control. Replication is paused when the replica queue exceeds this limit. In the default operation mode, flow control limit is dynamically recalculated based on the amount of nodes in the cluster, but this recalculation can be turned off with use of the gcs.fc_master_slave variable to make manual setting of the gcs.fc_limit having an effect (e.g., for configurations when writing is done to a single node in Percona XtraDB Cluster).

gcs.fc_master_slave

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	NO
Default Value:	NO

This variable is used to specify if there is only one source node in the cluster. It affects whether flow control limit is recalculated dynamically (when NO) or not (when YES).

gcs.max_packet_size

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	64500

This variable is used to specify the writeset size after which they will be fragmented.

gcs.max_throttle

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0.25

This variable specifies how much the replication can be throttled during the state transfer in order to avoid running out of memory. Value can be set to 0.0 if stopping replication is acceptable in order to finish state transfer.

gcs.recv_q_hard_limit

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	9223372036854775807

This variable specifies the maximum allowed size of the receive queue. This should normally be (RAM + swap) / 2. If this limit is exceeded, Galera will abort the server.

gcs.recv_q_soft_limit

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0.25

This variable specifies the fraction of the gcs.recv_q_hard_limit after which replication rate will be throttled.

gcs.sync_donor

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	No

This variable controls if the rest of the cluster should be in sync with the donor node. When this variable is set to YES, the whole cluster will be blocked if the donor node is blocked with SST.

gmcast.listen_addr

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	tcp://0.0.0.0:4567

This variable defines the address on which the node listens to connections from other nodes in the cluster.

gmcast.mcast_addr

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	None

This variable should be set up if UDP multicast should be used for replication.

gmcast.mcast_ttl

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	1

This variable can be used to define TTL for multicast packets.

gmcast.peer_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT3S

This variable specifies the connection timeout to initiate message relaying.

gmcast.segment

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable specifies the group segment this member should be a part of. Same segment members are treated as equally physically close.

gmcast.time_wait

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT5S

This variable specifies the time to wait until allowing peer declared outside of stable view to reconnect.

gmcast.version

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This variable shows which gmcast protocol version is being used.

ist.recv_addr

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	value of wsrep_node_address

This variable specifies the address on which the node listens for Incremental State Transfer (IST).

pc.announce_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT3S

Cluster joining announcements are sent every ½ second for this period of time or less if other nodes are discovered.

pc.checksum

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	true

This variable controls whether replicated messages should be checksummed or not.

pc.ignore_quorum

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	false

When this variable is set to TRUE, the node will completely ignore quorum calculations. This should be used with extreme caution even in source-replica setups, because replicas won’t automatically reconnect to source in this case.

pc.ignore_sb

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	false

When this variable is set to TRUE, the node will process updates even in the case of a split brain. This should be used with extreme caution in multi-source setup, but should simplify things in source-replica cluster (especially if only 2 nodes are used).

pc.linger

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT20S

This variable specifies the period for which the PC protocol waits for EVS termination.

pc.npvo

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	false

When this variable is set to TRUE, more recent primary components override older ones in case of conflicting primaries.

pc.recovery

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	true

When this variable is set to true, the node stores the Primary Component state to disk. The Primary Component can then recover automatically when all nodes that were part of the last saved state re-establish communication with each other. This feature allows automatic recovery from full cluster crashes, such as in the case of a data center power outage. A subsequent graceful full cluster restart will require explicit bootstrapping for a new Primary Component.

pc.version

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This status variable is used to check which PC protocol version is used.

pc.wait_prim

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	true

When set to TRUE, the node waits for a primary component for the period of time specified in pc.wait_prim_timeout. This is useful to bring up a non-primary component and make it primary with pc.bootstrap.

pc.wait_prim_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	PT30S

This variable is used to specify the period of time to wait for a primary component.

pc.weight

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	1

This variable specifies the node weight that’s going to be used for Weighted Quorum calculations.

protonet.backend

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	asio

This variable is used to define which transport backend should be used. Currently only ASIO is supported.

protonet.version

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	0

This status variable is used to check which transport backend protocol version is used.

repl.causal_read_timeout

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	PT30S

This variable specifies the causal read timeout.

repl.commit_order

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	3

This variable is used to specify out-of-order committing (which is used to improve parallel applying performance). The following values are available:

• 0 - BYPASS: all commit order monitoring is turned off (useful for measuring performance penalty)

• 1 - OOOC: allow out-of-order committing for all transactions

• 2 - LOCAL_OOOC: allow out-of-order committing only for local transactions

• 3 - NO_OOOC: no out-of-order committing is allowed (strict total order committing)

repl.key_format

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	Yes
Default Value:	FLAT8

This variable is used to specify the replication key format. The following values are available:

• FLAT8 - short key with higher probability of key match false positives

• FLAT16 - longer key with lower probability of false positives

• FLAT8A - same as FLAT8 but with annotations for debug purposes

• FLAT16A - same as FLAT16 but with annotations for debug purposes

repl.max_ws_size

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	2147483647

This variable is used to specify the maximum size of a write-set in bytes. This is limited to 2 gygabytes.

repl.proto_max

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	7

This variable is used to specify the highest communication protocol version to accept in the cluster. Used only for debugging.

socket.checksum

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	2

This variable is used to choose the checksum algorithm for network packets. The following values are available:

• 0 - disable checksum

• 1 - plain CRC32 (used in Galera 2.x)

• 2 - hardware accelerated CRC32-C

socket.ssl

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	No

This variable is used to specify if SSL encryption should be used.

socket.ssl_ca

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No

This variable is used to specify the path to the Certificate Authority (CA) certificate file.

socket.ssl_cert

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No

This variable is used to specify the path to the server’s certificate file (in PEM format).

socket.ssl_key

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No

This variable is used to specify the path to the server’s private key file (in PEM format).

socket.ssl_compression

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	Yes

This variable is used to specify if the SSL compression is to be used.

socket.ssl_cipher

Option	Description
Command Line:	Yes
Config File:	Yes
Scope:	Global
Dynamic:	No
Default Value:	AES128-SHA

This variable is used to specify what cypher will be used for encryption.