Search before asking

• I searched in the issues and found nothing similar.

Fluss version

main (development)

Please describe the bug 🐞

After an unclean TabletServer shutdown, log recovery reads
recovery-point-offset-checkpoint to decide where segment recovery should
start. Today this checkpoint is mainly written during clean shutdown or
truncate paths, so after an unclean shutdown it can be stale or missing.
When that happens, recovery may scan many already-flushed segments across
all replicas, which can significantly slow TabletServer startup.

Solution

No response

Are you willing to submit a PR?

• I'm willing to submit a PR!