Horton

Re: Server goes to Recovery Mode when run a SQL

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: PegoraroF10 <marcos@f10.com.br>
Cc: pgsql-general@postgresql.org
Date: 2020-08-12T22:09:02Z
Lists: pgsql-general 
PegoraroF10 <marcos@f10.com.br> writes:
> Hi Michael, so sorry for the delay. I did never had that problem again, so
> tougth it was solved, but seems not. We have two different SQL which pushs
> server to recovery mode. This SQL is the hardest one, because you´ll have to
> wait some time to get the error. Just runs this SQL, opens your HTOP and
> wait, you´ll see your memory growing and growing until ... 

Are you sure this query ever terminates?  It looks to me like it just
keeps producing tuples indefinitely until you kill it or something runs
out of memory.

If the response to "out of memory" is a server crash rather than something
nicer, you probably need to corral your OOM killer.  See

https://www.postgresql.org/docs/current/kernel-resources.html#LINUX-MEMORY-OVERCOMMIT

			regards, tom lane

Commits

  1. Be more careful about the shape of hashable subplan clauses.