Hi Mahi,
Are you trying to upgrade your Hazelcast pods one-at-a-time in a rolling fashion or are you taking an outage, upgrading the images for every pod in the cluster, and then bringing the cluster back online? If you're using Hazelcast open source edition, a rolling restart is not supported (rolling restart is a feature of Hazelcast Enterprise edition only).
If you're using open source, you can still do the upgrade, you just can't have mixed versions of Hazelcast in a single cluster. For example, you can't have a cluster with 7 nodes, 2 of which are running 5.1.3 and and 5 of which are running 5.0.2.
Instead, you have to bring the entire cluster down (note that this will mean an application outage, so you may want to do it during a low-traffic period) upgrade all 7 nodes to 5.1.3 (which for you probably means upgrading the pod container images) and then bring the cluster back up.
Note that this limitation is not present in Hazelcast Enterprise Edition, will which allow you to do a rolling upgrade without taking an outage.
Take care,
Lucas
Lucas Beeler | Principal Architect, Office of the CTO | lucas....@hazelcast.com | | 3000 El Camino Real, Bldg 4, Ste 200 | Palo Alto, CA 94306 | USA +1 (650)395-7716 | http://www.hazelcast.com |
|
|