SEZ9 commented on PR #12678:
URL: https://github.com/apache/seatunnel/pull/12678#issuecomment-6992336759

   @kalayciburak thanks for the updates. Here is where the two points stand.
   
   **Restore seeding / bound across restarts:** this is settled for this PR. 
Seeding the deque from storage stays in #12094. Your description update, as you 
describe it, covers the limitation: the new coordinator starts with an empty 
deque, doesn't select old files, and repeated restarts can exceed 2x. Nothing 
more is needed here.
   
   **Delete failure during pruning:** this is still open. The behaviour you 
describe at `71ac656652f0` is the right direction: ids go back to the front in 
order when `deleteCheckpoint` reports an `Exception`, and they are retried on 
the next eligible completion. But as you noted, 
`CheckpointCoordinatorTest#testRetentionDeleteFailureRetriesOnNextCompletion` 
mocks the delete failures and stubs `notifyCompleted`, so it doesn't exercise a 
storage failure end to end.
   
   Remaining asks:
   
   1. Push the storage-backed test, showing that:
      - checkpoint completion still succeeds after a reported deletion failure, 
and
      - a later completion actually removes the id that failed.
   2. Link a Build run for the exact revision that contains that test.
   
   Once both are in, I'll take another look.
   
   <!-- streview-comment:2438 -->


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to