Bugzilla – Bug 4356
Function of skipCount across restart
Last modified: 2013-01-16 15:58:10 UTC
At first I assumed there was nothing to say here.. the skip count(s) would get reset to '0' and we'd start counting against skip-limit all over again.
I realized I'm making an assumption though... that it's correct to reset the three "skip metrics" for StepExecution to 0.
If we were to try to persist these values across restart, we have another subtlety:
We can double count skips against certain records if we:
b) hit a rollback within the same checkpoint
c) skip the same record on restart
I think that assuming we do reset the count to 0, it is fine to "recount"...but it's more in the case where we try to persist the count that double-counting is an issue.
The skip count is per job execution.