Hi all,
I am facing very unusual behaviour in our production postgres system running on pg15 and deployed on AWS.
pgwal is mounted on a separate EBS drive with size of 100 GBs (on disk size as seen from df -hT).
max_wal_size = 95 GB, min_wal_size = 1 GB, checkpoint_completion_target = 0.9
Issue is that pgwal size increased beyond 95 GB limits, and reached max drive capacity of 100 GBs while ingesting huge data via copy insert.
Although the root cause of the problem to me looks like ebs volume size is too close to max wal size and by the time limit was breached and check point started, it was too late as WAL writing continued to happen and reached max limit of drive, would like to know if someone has any ideas of this.
Also, should we add a suggestion in doc regarding disk space and max wal size as max wal size looks more like an upper limit which WAL is unlikely to surpass(by large margin).
Happy to revert with more details if required.
Thanks,
Ankit