Re: BUG #17977: PorstGreSQL in a jail crashes randomly with Signal 10 bus error - Mailing list pgsql-bugs

From Cory Albrecht
Subject Re: BUG #17977: PorstGreSQL in a jail crashes randomly with Signal 10 bus error
Date
Msg-id CAMW5rYKWncKWroaKSyPiTnX=n=4-0kSegwG4V_Z-4BcATtTjEg@mail.gmail.com
Whole thread Raw
In response to Re: BUG #17977: PorstGreSQL in a jail crashes randomly with Signal 10 bus error  (Tom Lane <tgl@sss.pgh.pa.us>)
Responses Re: BUG #17977: PorstGreSQL in a jail crashes randomly with Signal 10 bus error  (Tom Lane <tgl@sss.pgh.pa.us>)
List pgsql-bugs
After things percolated overnight, and realising that scripts O had made to try and overload the server had connected with SSL, not Keberos, encryption, I decided to turn off the GSS/Kerberos settings so that everything connected only with SSL encryption. So far today in working with QGIS I have no SIGBUS crashes. Perhaps that reminds anybody of any previous issues?

I've currently got my poudriere machine rebuilding the the package (I needed to upgrade the servers from 14.6 to 14.8 anyways) to get a coredump, but that will take some time.

On Thu, Jun 15, 2023 at 3:07 PM Tom Lane <tgl@sss.pgh.pa.us> wrote:
Cory Albrecht <coryca@gmail.com> writes:
> Also, which PID do I attach GDB to?
> If it's the individual database ones like 15947 or 15990 in the
> above example, I have no way of knowing which one will get a SIGBUS. Is the
> main process (e.g. 15892) OK?

If you don't know that, the best way is to collect a core file and
then gdb that.  Starting the postmaster under "ulimit -c unlimited"
ought to be enough.

                        regards, tom lane

pgsql-bugs by date:

Previous
From: Tom Lane
Date:
Subject: Re: Server closed the connection unexpectedly
Next
From: Tom Lane
Date:
Subject: Re: BUG #17979: (pg_tsparser + ts_headline) duplicate output if dot present