repmgr standby promote

repmgr standby promote — promote a standby to a primary

Description

Promotes a standby to a primary if the current primary has failed. This command requires a valid repmgr.conf file for the standby, either specified explicitly with -f/--config-file or located in a default location; no additional arguments are required.

Important

If repmgrd is active, you must execute repmgr service pause (repmgr 4.2 - 4.4: repmgr service pause) to temporarily disable repmgrd while making any changes to the replication cluster.

If the standby promotion succeeds, the server will not need to be restarted. However any other standbys will need to follow the new primary, and will need to be restarted to do this.

Beginning with repmgr 4.4, the option --siblings-follow can be used to have all other standbys (and a witness server, if in use) follow the new primary.

Note

If using repmgrd, when invoking repmgr standby promote (either directly via the promote_command, or in a script called via promote_command), --siblings-follow must not be included as a command line option for repmgr standby promote.

In repmgr 4.3 and earlier, repmgr standby follow must be executed on each standby individually.

repmgr will wait for up to promote_check_timeout seconds (default: 60) to verify that the standby has been promoted, and will check the promotion every promote_check_interval seconds (default: 1 second). Both values can be defined in repmgr.conf.

Note

If WAL replay is paused on the standby, and not all WAL files on the standby have been replayed, repmgr will not attempt to promote it.

This is because if WAL replay is paused, PostgreSQL itself will not react to a promote command until WAL replay is resumed and all pending WAL has been replayed. This means attempting to promote PostgreSQL in this state will leave PostgreSQL in a condition where the promotion may occur at a unpredictable point in the future.

Note that if the standby is in archive recovery, repmgr will not be able to determine if more WAL is pending replay, and will abort the promotion attempt if WAL replay is paused.

Example

      $ repmgr -f /etc/repmgr.conf standby promote
      NOTICE: promoting standby to primary
      DETAIL: promoting server "node2" (ID: 2) using "pg_ctl -l /var/log/postgres/startup.log -w -D '/var/lib/postgres/data' promote"
      server promoting
      NOTICE: STANDBY PROMOTE successful
      DETAIL: server "node2" (ID: 2) was successfully promoted to primary

User permission requirements

pg_promote() (PostgreSQL 12 and later)

From PostgreSQL 12, repmgr will attempt to use the built-in pg_promote() function to promote a standby to primary.

By default, execution of pg_promote() is restricted to superusers. If the repmgr user does not have permission to execute pg_promote(), repmgr will fall back to using "pg_ctl promote".

Tip

Execute repmgr standby promote with the --dry-run to check whether the repmgr user has permission to execute pg_promote().

If the repmgr user is not a superuser, execution permission for this function can be granted with e.g.:

    GRANT EXECUTE ON FUNCTION pg_catalog.pg_promote TO repmgr

Note that permissions are only effective for the database they are granted in, so this must be executed in the repmgr database to be effective.

For more details on pg_promote(), see the PostgreSQL documentation.

Options

--dry-run

Check if this node can be promoted, but don't carry out the promotion.

--siblings-follow

Have all sibling nodes (nodes formerly attached to the same upstream node as the promotion candidate) follow this node after it has been promoted.

Note that a witness server, if in use, is also counted as a "sibling node" as it needs to be instructed to synchronise its metadata with the new primary.

Important

Do not provide this option when configuring repmgrd's promote_command.

-F
--force

Ignore warnings and continue anyway.

This option is relevant in the following situations if --siblings-follow was specified:

  • If one or more sibling nodes was not reachable via SSH, the standby will be promoted anyway.
  • If the promotion candidate has insufficient free walsenders to accomodate the standbys which will be attached to it, the standby will be promoted anyway.
  • If replication slots are in use but the promotion candidate has insufficient free replication slots to accomodate the standbys which will be attached to it, the standby will be promoted anyway.

Note that if the -F/--force option is used when any of the above situations is encountered, the onus is on the user to manually resolve any resulting issues.

Configuration file settings

The following parameters in repmgr.conf are relevant to the promote operation:

  • promote_check_interval: interval (in seconds, default: 1 second) to wait between each check to determine whether the standby has been promoted.

  • promote_check_timeout: time (in seconds, default: 60 seconds) to wait to verify that the standby has been promoted before exiting with ERR_PROMOTION_FAIL.

  • service_promote_command: a command which will be executed instead of pg_ctl promote or (in PostgreSQL 12 and later) pg_promote().

    This is intended for systems which provide a package-level promote command, such as Debian's pg_ctlcluster, to promote the PostgreSQL from standby to primary.

Exit codes

Following exit codes can be emitted by repmgr standby promote:

SUCCESS (0)

The standby was successfully promoted to primary.

ERR_DB_CONN (6)

repmgr was unable to connect to the local PostgreSQL node.

PostgreSQL must be running before the node can be promoted.

ERR_PROMOTION_FAIL (8)

The node could not be promoted to primary for one of the following reasons:

  • there is an existing primary node in the replication cluster
  • the node is not a standby
  • WAL replay is paused on the node
  • execution of the PostgreSQL promote command failed

Event notifications

A standby_promote event notification will be generated.