[tidb] PD failover count throws errors even when there is no error

[unw9527]

What problem does this PR solve?

When failovering a broken PD member, this line gives us an error even if no PD member fails (i.e. pdDeletedFailureReplicas=0). When we inspect the error log, we find that the error message "PD failover replicas (0) reaches the limit (0), skip failover" is constantly written to the operator log and is very confusing to user. However, there is actually no error if no PD member fails. We believe when pdDeletedFailureReplicas=0, the operator should not throw the error message "PD failover replicas (0) reaches the limit (0), skip failover".

This PR provides a straightforward fix for this problem.

What is changed and how does it work?

We fix it by adding one more condition in the if statement in this line to check whether pdDeletedFailureReplicas is larger than 0, as illustrated below:

if pdDeletedFailureReplicas >= *tc.Spec.PD.MaxFailoverCount && pdDeletedFailureReplicas > 0 {
	klog.Errorf("PD failover replicas (%d) reaches the limit (%d), skip failover", pdDeletedFailureReplicas, *tc.Spec.PD.MaxFailoverCount)
	return nil
}

Code changes

• Has Go code change
• Has CI related scripts change

Tests

• Unit test
• E2E test
• Manual test
• No code

This is a simple fix and we suppose no test above is needed.

Side effects

• Breaking backward compatibility
• Other side effects:

Related changes

• Need to cherry-pick to the release branch
• Need to update the documentation

Release Notes

Please refer to Release Notes Language Style Guide before writing the release note.

NONE

[unw9527]

PR created: pingcap/tidb-operator#4608

[tianyin]

@unw9527 you are a PR expert now.

[unw9527]

Confirmed: pingcap/tidb-operator#4608 (review)

[tylergu]

Fixed by pingcap/tidb-operator#4608