When upgrading Nexus-RM Image versions, a DB-Schema upgrade is scheduled that may take several minutes to complete. If we relied on the default liveness-probe configuration, it is likely that the container will be terminated before this DB transition is complete. Our current solution is to increase the Value of statefulset.livenessProbe.initialDelaySeconds to something like 30 minutes. The side effect of such a setting is, that when a normal reboot occurs and no DB migration is required, we would still have a timeframe of 30 minutes before the probe gets active. A better approach would be to get rid of the initialDelaySeconds values in the liveness and readiness probe and introduce a startup-probe with a default setting of:
When upgrading Nexus-RM Image versions, a DB-Schema upgrade is scheduled that may take several minutes to complete. If we relied on the default liveness-probe configuration, it is likely that the container will be terminated before this DB transition is complete. Our current solution is to increase the Value of
statefulset.livenessProbe.initialDelaySeconds
to something like 30 minutes. The side effect of such a setting is, that when a normal reboot occurs and no DB migration is required, we would still have a timeframe of 30 minutes before the probe gets active. A better approach would be to get rid of theinitialDelaySeconds
values in the liveness and readiness probe and introduce a startup-probe with a default setting of: