Hi,
PG: 9.6.11
OS: CentOS
Env: AWS EC2
I've faced the following exceptions in PostgreSQL server log:
> could not truncate directory "pg_serial": apparent wraparound
Sometimes it repeats every 5 min and the longest period was 40 min.
In fact, I can't find any suspicious events happening that periods.
pg_wait_sampling didn't catch any events, no long queries (more than 60s), Autovacuum workers or transactions in "idle in transaction" state were in action at this time.
>You should not see the errors you are reporting nor
>the warning I mentioned unless a serializable transaction remains
>active long enough for about 1 billion transaction IDs to be
>consumed.
Database age now is just 18.5 millions of transactions.
Server has two standbys (sync and async), hot_standby_feedback is off.
Please advice what I can do to find a reason of these exceptions.