A backup taken on a standby instance can complete successfully (CNPG Backup reports completed, no error) while never being relayed to tier2. kopia snapshot list --all against the tier2 repository shows nothing for the affected cluster. There is no error, warning, or metric indicating the relay was skipped.
Starting a backup preferring the standby (which is the default) on the tier2 cluster and server samples is enough to reproduce the issue, for example using the cnpg plugin:
kubectl cnpg backup cluster-example -m plugin --plugin-name klio.cnpg.io
A backup taken on a standby instance can complete successfully (CNPG Backup reports completed, no error) while never being relayed to tier2. kopia snapshot list --all against the tier2 repository shows nothing for the affected cluster. There is no error, warning, or metric indicating the relay was skipped.
Starting a backup preferring the standby (which is the default) on the tier2 cluster and server samples is enough to reproduce the issue, for example using the cnpg plugin: