Re: Debugging a backend stuck consuming CPU

From: Merlin Moncure <mmoncure(at)gmail(dot)com>
To: "ktm(at)rice(dot)edu" <ktm(at)rice(dot)edu>
Cc: PostgreSQL General <pgsql-general(at)postgresql(dot)org>
Subject: Re: Debugging a backend stuck consuming CPU
Date: 2016-05-19 14:30:15
Message-ID: CAHyXU0xNJ_UTu9AXWRWwbN_VCr7D0jBMCoEa2U_LL1Q_k1=g2w@mail.gmail.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-general

On Thu, May 19, 2016 at 8:46 AM, ktm(at)rice(dot)edu <ktm(at)rice(dot)edu> wrote:
> Hi PostgreSQL community,
>
> I am investigating a problem with a backend that appears to be stuck
> and spinning while performing a "DISCARD ALL" command. The system is
> running an older release 9.2.2. Are there any bugs that could be
> causing this behavior? How can I tell what the process is actually
> doing? It does not respond to either pg_cancel_backend() or to
> pg_terminate_backend() so it is acting like it is not receiving the
> signals. Any ideas on how to debug this would be appriciated.

1. Grab a stack trace and a 'perf top' for forensic analysis in case
it happens again
2. Install latest 9.2 binaries
3. Restart database

There's not much value in analyzing such an out of date database;
there's a very high probability your bug has already been fixed.

merlin

In response to

Browse pgsql-general by date

  From Date Subject
Next Message Peter Juhasz 2016-05-19 14:37:48 PQcancel may hang in the recv call
Previous Message Merlin Moncure 2016-05-19 14:19:14 Re: PG wire protocol question