K8SPG-217: Fix missing primary deployment after restore

[egegunes]

The problem only occurs if primary is failed over to a replica before the restore.

The operator stores the current primary in an annotation of pgcluster (i.e. current-primary: cluster1). After the bootstrap finishes, operator creates the primary deployment using this value as the deployment name. After the primary is created, it proceeds to create replica deployments.

If a failover is performed and a replica deployment is promoted to primary, the operator updates the annotation (i.e. current-primary: cluster1-repl2).

If a restore is started at this moment in the cluster:

1. Bootstrap finishes
2. Operator creates cluster1-repl2 as the primary deployment
3. Operator creates Pgreplica objects, cluster1-repl1 is created but cluster1-repl2 is skipped since it already exists

This commit updates the current-primary annotation with cluster name just before the restore, so restore is performed on PVC cluster1 and replicas are created as expected.

[JNKPercona]

Test name	Status
demand-backup	passed
scheduled-backup	passed
version-service	passed
init-deploy	passed
upgrade	passed
smart-update	passed
scaling	passed
recreate	passed
affinity	passed
monitoring	passed
self-healing	passed
operator-self-healing	passed
clone-cluster	passed
tls-check	passed
users	passed
ns-mode	passed
We run 16 out of 16

commit: https://github.com/percona/percona-postgresql-operator/pull/295/commits/5516942628ea4933f2d827cbda1c297257b5c6ef image: perconalab/percona-postgresql-operator:PR-295-55169426-pgo-apiserver

image: perconalab/percona-postgresql-operator:PR-295-55169426-pgo-event

image: perconalab/percona-postgresql-operator:PR-295-55169426-pgo-rmdata

image: perconalab/percona-postgresql-operator:PR-295-55169426-pgo-scheduler

image: perconalab/percona-postgresql-operator:PR-295-55169426-postgres-operator

image: perconalab/percona-postgresql-operator:PR-295-55169426-pgo-deployer