Haolin Wang
Haolin Wang
> Is it proved that SO_REUSEADDR makes no difference for outgoing connections? Or did I test it wrongly? I think needs to consume up all outgoing ports (maybe by narrowing...
"Cannot assign requested address" is EADDRNOTAVAIL.
So that indicates SO_REUSEADDR doesn't take effect at outgoing TCP requests, right ? I think we don't need this change based on the current observation until we do see some...
Just a reminder that the commit message also needs to be filled with detail description of the change before merging.
This is trying to revise PR https://github.com/greenplum-db/gpdb/pull/10963
> We introduced another Bitmap to improve the efficiency of checking the visibility of a tuple, can we apply this to the normal scan? Such as SeqScan (in `aocs_getnext()`). >...
This is separate the first commit out of the PR https://github.com/greenplum-db/gpdb/pull/12746 for being ready to be integrated. A specific test case was also added.
On 6X, if TRUNCATE interleaves between END_OF_RECOVERY and its following CHECKPOINT, after COMMIT_PREPARED, will unlink the old AO segment file but the later CHECKPOINT will queue and apply the fsync...
Reopen it, it is a different issue, #12699 cannot fix it.
More investigation, primary performs insert fsync for AO segment via backend other than checkpointer, while mirror handles insert fsync during redo by enqueueing fsync request to checkpointer. The problem happends...