Pods are restarted when new shards/replicas are added

[xiedeyantu]

release 0.12.0 first improvement:
Pods are no longer restarted when new shards/replicas are added.
But it will be restarted until 0.13.5.
Because of some cluster scope label changes, pods are restarted.

clickhouse.altinity.com/chiScopeCycleOffset: "3"
clickhouse.altinity.com/chiScopeIndex: "3"
clickhouse.altinity.com/clusterScopeCycleOffset: "3"
clickhouse.altinity.com/clusterScopeIndex: "3"

Is there any way to fix this problem？

[alex-zaitsev]

Hi @xiedeyantu, could you describe what you are doing in more detail? CHI would be helpful. It looks like you are using CircularReplication, but we do not understand how it can change the cycle index and offset for existing pods.

If there is still a problem, the proper fix would be to remove those labels completely, they are just for the reference.

[xiedeyantu]

Hi @xiedeyantu, could you describe what you are doing in more detail? CHI would be helpful. It looks like you are using CircularReplication, but we do not understand how it can change the cycle index and offset for existing pods.

If there is still a problem, the proper fix would be to remove those labels completely, they are just for the reference.

I have come up with a solution to this problem. The method is to patch the labels of STS first, and then update STS. I have submitted a pr #667, could you take a look at this pr?

[alex-zaitsev]

Fixed in https://github.com/Altinity/clickhouse-operator/releases/tag/0.14.0