trigger Cluster(s) upgrade when upgrading the Plugin

[l00ptr]

Upgrading the plugin does not actually trigger an upgrade of the clusters using it. To achieve this, we would probably need to write a custom controller to detect such changes and enforce reconciliation.

I think it's also an issue with the Barman plugin. More information can be found here.

[fguiet]

Hi,

Actually, I just did a plugin upgrade from v0.0.2 to v0.0.3 and it did trigger a restart of the clusters.
The only thing I had to do was to update the plugin config to add the new exporter config :

exporterConfig:
    enabled: true
    resourcesRequirement:
      requests:
      cpu: 1m
      memory: 1Mi
      limits:
        cpu: 500m
        memory: 512Mi

Then kill the cluster pod so the exporter sidecar starts

[AnizR]

Hello,
The automatic trigger of cluster restart when upgrading a plugin has been introduced in cnpg 1.29.2 (PR: cloudnative-pg/cloudnative-pg#10836).
Therefore, if your controller is not up to date, you might not have your cluster that automatically restarts.