Specify parameters in the format described below.

After execution of FEPAction CR, status is reflected in fepStatus field that is dynamically inserted in current FEPAction CR as needed.

fepStatus field used for FEPAction CR are described here

apiVersion: fep.fujitsu.io/v1
kind: FEPAction
fepActionStatus:
 fepActionCondition: Success
metadata:
 name: new-fep-reload-action
 namespace: myns
spec:
 fepAction:
  args:
  - new-fep-sts-0
  - new-fep-sts-1
  type: reload
 sysExtraLogging: false
 targetClusterName: new-fep

The reload action will manually reload the FEP database on the targeted FEPCluster.

"reload" action type expects users to specify the name of individual FEP pods that they want to run the database reload operation on. They specify that in the args section under the FEPAction CR spec as below :

spec:
fepAction:
 args:
 - nf-131851-sts-0
 - nf-131851-sts-1
 type: reload
 targetClusterName: nf-131851

The restart action will manually restart the FEP database on the targeted FEPCluster.

"restart" action type expects users to specify the name of individual FEP pods that they want to run the database restart operation on. They specify that in the args section under the FEPAction CR spec as below:

spec:
fepAction:
 args:
 - nf-131851-sts-0
 - nf-131851-sts-1
 type: restart
 targetClusterName: nf-131851

The pod_restart action will restart specified list of POD for given target cluster. User can specify key word ‘ALL’ under ‘args’ section to restart all pods in target cluster. Alternatively, user can give the list of pods to be started in target cluster. User should either give ALL or the list of the pods.

This action restarts the replica pods first. Once all replicas have been restarted, it switches over the mastership to one of the replica before restarting old master pod. If it is a single node cluster, master will be restarted in its current state. This action is automatically created to restart pods when image or machine specs are changed for fep or backup container depending on autoPodRestart flag in FEPCluster CR (Refer to "1.2.1 FEPCluster Custom Resource Parameters" for details.):

spec:
fepAction:
args:
 - nf-131851-sts-0
 - nf-131851-sts-1
 type: pod_restart
 targetClusterName: nf-131851

The list action will return the status of the targeted FEPCluster.

"list" action type expects users to specify just the target cluster name to list the details of the same. Looks like below:

spec:
fepAction:
 type: list
 targetClusterName: nf-131851

The switchover action performs a manually switchover of the current leader/primary database from one pod to another pod of the targeted FEPCluster.

"switchover" action type expects users to specify the name of the target cluster that they want to perform switchover. args section is not required for switchover as FEPAction operator code will internally find it and promote new master. FEPAction CR spec as below:

spec:
fepAction:
 type: switchover
 targetClusterName: nf-131851

The failover action performs a manually failover of the current primary database from one pod to another pod of the targeted FEPCluster. The difference between switchover and failover is that, switchover expects the primary database is running at the time whereas failover can force switchover of primary role from a non-responding pod to another pod. Note that failover is a disruptive action and may cause data lost.

"failover" action type expects users to specify the names of the candidate pods that they want to failover to. They specify that in the args section under the FEPAction CR spec as below:

spec:
fepAction:
 args:
 - nf-131851-sts-1
 - nf-131851-sts-2
 type: failover
 targetClusterName: nf-131851

Here, nf-131851-sts-1 and nf-131851-2 are the candidate pods to failover to. In this example, the current primary pod would be nf-131851-sts-0.

"pgpool2_restart" action type expects users to specify the name of individual FEPPgpool2 resource that they want to restart operation on. They specify that in the targetPgpool2Name section under the FEPAction CR spec as below:

spec:
fepAction:
 type: pgpool2_restart
 targetPgpool2Name: nf-131851-pgpool2

The "backup" action performs a backup on the target FEPCluster.

The "backup" action type requires you to specify the type of backup and the repository in which to store the data.

In the fepAction section of the FEPAction custom resource specification, specify the following:.

spec: 
  targetClusterName: new-fep 
  fepAction:  
    type: backup 
    backupType: full 
    backupRepo: 1

The open_tde_masterkey action opens a keystore for a TDE-enabled target cluster.

The "open_tde_masterkey" action type requires the user to specify the name of the target cluster on which the keystore will be opened. The args section is not required.

Specify the following:

spec:
  targetClusterName: nf-131851
  fepAction:
    type: open_tde_masterkey

The create_extension action executes "CREATE EXTENSION" on the target FEPCluster and installs the extension.

In fepAction.args, specify the "extension name, version", "database", "schema", and "apply CASCADE options" to be installed.

An example specification is shown below.

spec: 
  targetClusterName: new-fep
  fepAction: 
    type: create_extension
    args: 
      extension: "vci"
      version: "2.0"
      database: "mydb"
      schema: "public"
      cascade: "true"

The alter_extension action executes "ALTER EXTENSION" on the target FEPCluster to change the definition of the extension.

You can execute ALTER EXTENSION on the specified FEPCluster container by applying fepAction.args.version with the version to be updated.

Here is an example of a FEPAction custom resource that updates the version of pg_stat_statements:. When a custom resource is applied, the version of pg_stat_statements is updated to the specified version.

spec:
  targetClusterName: new-fep
  fepAction:
    type: alter_extension
    args:
      extension:pg_stat_statements
      version:1.8
      database:sample

The update_admin_password action redefines the password for SUPERUSER "postgres" on the target FEPCluster with a random value.

This action will be executed when the FEPCluster custom resource spec.fepChildCrVal.sysUsers.pgAdminPassword is not defined.

An example specification is shown below.

spec: 
fepAction: 
type: update_admin_password
targetClusterName: new-fep

You can run the "pgbackrest expire" command on the FEPPod to remove expired backups.

The "pgbackrest expire" command is normally run automatically upon a successful backup, but it can be run by the user, for example, when the definition of the number of generations to retain for a backup is reduced, and the backup data can be deleted so that the number of retained generations conforms to the changed definition.

If you want to reduce the number of backup retention generations and free up disk space, apply "backup_expire" in the FEPAction after changing the retention setting for backup data under the FEPCluster custom resource fepChildCrVal.backup.pgbackrestParams.

You can specify the repository from which to remove the backup by specifying args.repo.

The following is an example of changing the retention setting for backup data in a FEPCluster custom resource.

You want to reduce the number of backup generations stored in S3.

spec: 
  fepChildCrVal:
    backup:
      pgbackrestParams:
        repo2-type=s3
        repo2-retention-full=5   # Change it to the number of generations you want to keep
        repo2-retention-full-type=time

The following is an example of a FEPAction custom resource that reduces the number of backup generations:

Since the backup repository for s3 is specified as repo2-type in pgbackrestParams, specify 2 for spec.fepActoin.repo.

apiVersion: fep.fujitsu.io/v1 
kind: FEPAction
metadata: 
  name: backup-expire-action
spec: 
  targetCluster: new-fep
  fepAction:
    type: backup_expire
    args:
      repo: 2

promote_standby promotes the FEP database in the disaster recovery environment from Standby DB to Primary DB. You must specify the DB cluster to be promoted.

The following shows a specification example.

spec:
  fepAction:
    type: promote_standby
  targetClusterName: my-fep

The update_multi_master_replication action updates the bidirectional replication definition according to changes in the bidirectional replication definition specified in the ConfigMap referenced by the spec.fep.multiMasterReplication.configMapName field of the target FEPCluster custom resource.

The fixed_stats action performs a statistics operation (pg_dbms_stats) on the target FEPCluster.

The fixed_stats action must specify the operations that can be performed by pg_dbms_stats. Specify fixedStatsType: < operation type > in spec.fepAction.args. Depending on the operation type, there are additional required and optional arguments. Refer to the table below.

The following shows a specification example.

apiVersion: fep.fujitsu.io/v1
kind: FEPAction
metadata:
  name: fep-action-firxdStats
spec:
  fepAction:
    targetClusterName: new-fep
    type: fixed_stats 
    args:
      fixedStatsType: backup
      targetDb: mydb
      fixedObject: schema.table.attaname
      comment: "xxyyzz"