| From: | Michael Paquier <michael(at)paquier(dot)xyz> |
|---|---|
| To: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | Thunder <thunder1(at)126(dot)com>, Kuntal Ghosh <kuntalghosh(dot)2007(at)gmail(dot)com>, PostgreSQL Hackers <pgsql-hackers(at)lists(dot)postgresql(dot)org> |
| Subject: | Re: PANIC :Call AbortTransaction when transaction id is no normal |
| Date: | 2019-05-14 03:37:39 |
| Message-ID: | 20190514033739.GH1418@paquier.xyz |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Mon, May 13, 2019 at 11:28:51PM -0400, Tom Lane wrote:
> OK, that's fair. The SIG_DFL change I suggested will fix that problem
> for SIGINT etc (except SIGQUIT, for which you should be *expecting*
> a core file). I agree with Michael that we do not wish to change what
> happens for an internal error; but external signals do not represent
> a bug in PG, so forcing a PANIC for those seems unwarranted.
No objections from here to change the signal handlers. Still, I would
like to understand why the bootstrap process has been signaled to
begin with, particularly for an initdb, which is not really something
that should happen on a server where an instance runs. If you have a
too aggressive monitoring job, you may want to revisit that as well,
because it is able to complain just with an initdb.
--
Michael
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michael Paquier | 2019-05-14 03:50:09 | Re: [HACKERS] Unlogged tables cleanup |
| Previous Message | Tom Lane | 2019-05-14 03:29:59 | Re: VACUUM can finish an interrupted nbtree page split -- is that okay? |