5569 updates volume-snapshot restore process to fail the entire restore if any warmup job failed. We use this to quickly check the viability of restores and terminate restore processing early if a corruption is detected.

5572 updates volume-snapshot restore process to enable recovery from a corruption in a single TiKV through manual cluster operations. We use this in a full restore in case we encounter a corruption during this process.

These features are in conflict w/ each other. If we want to perform a full restore and use single TiKV recovery in the event of corruption, we cannot fail the restore during warmup and instead need to complete warmup stage and progress to restarting TiKVs. If we only want to check the viability of a restore, we are ok w/ failing the restore and not progressing to any further steps. Thus, we gate this failure behavior only behind the check-wal-only strategy.

What is changed and how does it work?

Gate restore failure on warmup failure only for check-wal-only warmup strategy.

Code changes

[x] Has Go code change
[ ] Has CI related scripts change

Tests

[x] Unit test
[ ] E2E test
[ ] Manual test
[ ] No code

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.

pingcap / tidb-operator

Fail restore if warmup fails only during `check-wal-only` strategy (#5621) #5636

What problem does this PR solve?