crash master after failover can cause unexpected downscaling

[xiaozhuang-a]

What version of redis operator are you using?

kubectl logs <_redis-operator_pod_name> -n <namespace>

redis-operator version: v0.18.1

Does this issue reproduce with the latest release?

yes

What operating system and processor architecture are you using (kubectl version)?

kubectl version Output

$ kubectl version

What did you do?

What did you expect to see?

cluster failure will not cause expansion or contraction of capacity

What did you see instead?
after the failover, the status of the original faulty master in the cluster nodes will be 'master fail', which will trigger downscaling

[xiaozhuang-a]

I think it should not be done through k8sutils The comparison between CheckRedisNodeCount and clusterSize determines whether to scale or not. It should be determined by comparing the current number of statefulset replicas with clusterSize, which can mask the calculation shard nums errors caused by intermediate states within the Redis cluster

[xiaozhuang-a]

If the design is feasible, I would like to propose a PR fix

[xiaozhuang-a]

option2: scaling after statefulset is ready