Re: trivial patch: show SIREAD pids in pg_locks

From: "Kevin Grittner" <Kevin(dot)Grittner(at)wicourts(dot)gov>
To: "Dan Ports" <drkp(at)csail(dot)mit(dot)edu>,"Jeff Davis" <pgsql(at)j-davis(dot)com>
Cc: <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: trivial patch: show SIREAD pids in pg_locks
Date: 2011-04-01 21:17:03
Message-ID: 4D95FA7F020000250003C171@gw.wicourts.gov
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Jeff Davis <pgsql(at)j-davis(dot)com> wrote:
> On Fri, 2011-04-01 at 13:00 -0400, Dan Ports wrote:
>> While looking into a SSI bug, I noticed that we don't actually
>> display the pid of the holding transaction, even though we have
>> that information available.
>
> Is there a chance that the PID will reference a backend that has
> either terminated or is idle?

Yes to both.

> That might be confusing.

With a barely larger patch we could suppress that, if that's
desirable. Of course, there is already the same issue for the
virtualtransaction column.

As Dan mentioned, it won't show for locks which are summarized using
SLRU, nor will it show for prepared transactions pending final
commit -- at least if they're loaded from disk after recovery. (I'm
not sure without some digging about a transaction which is prepared
and still pending commit if the server is still running from the
point of prepare.)

-Kevin

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Bruce Momjian 2011-04-01 21:48:16 Re: Bug in autovacuum.c?
Previous Message Jeff Davis 2011-04-01 21:02:59 Re: trivial patch: show SIREAD pids in pg_locks