Re: BUG #13970: Vacuum hangs on particular table; cannot be terminated - requires `kill -QUIT pid`

From: Brian Ghidinelli <brian(at)vfive(dot)com>
To: Alvaro Herrera <alvherre(at)2ndquadrant(dot)com>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #13970: Vacuum hangs on particular table; cannot be terminated - requires `kill -QUIT pid`
Date: 2016-03-12 22:06:31
Message-ID: C3F15AAD-8CFE-40C2-833B-C02624230DBE@vfive.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs


> On Mar 7, 2016, at 08:21, Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> wrote:
>
> Brian Ghidinelli wrote:
>> Any chance that a DDL modification like adding a column would effect an update of those values across the table?
>
> If a column is added without a default value, then the existing tuples
> are not touched. If a default value is added, then the complete table
> is rewritten. I suppose that the rewritten table would not include old
> Multixact values in XMAX, but I haven't checked.
>
> I wonder if it's possible for a column to be added in an operation that
> does not rewrite the table yet the relminmxid is advanced past the
> oldest multixact present.
>
> I assume, given the lack of traffic in this thread, that you have solved
> your problem already?

Thanks for the follow up Alvaro. I’m no longer on the bugs list so don’t know if this will make it through, but I went the route of recreating the table and updating all of the foreign key references and shared that on Feb 19th.

This seems to have fixed the problem with vacuuming the table however it did not update the multixact value which still reads 1. Any idea if this will still cause problems in the future?

# /usr/pgsql-9.4/bin/pg_controldata ~postgres/9.4/
pg_control version number: 942
Catalog version number: 201409291
Database system identifier: 6123001933830237052
Database cluster state: in production
pg_control last modified: Sat 12 Mar 2016 01:57:08 PM PST
Latest checkpoint location: 13A/54469E58
Prior checkpoint location: 13A/54129CC8
Latest checkpoint's REDO location: 13A/543F2908
Latest checkpoint's REDO WAL file: 000000010000013A00000054
Latest checkpoint's TimeLineID: 1
Latest checkpoint's PrevTimeLineID: 1
Latest checkpoint's full_page_writes: on
Latest checkpoint's NextXID: 0/92143770
Latest checkpoint's NextOID: 5609664
Latest checkpoint's NextMultiXactId: 84435
Latest checkpoint's NextMultiOffset: 167731
Latest checkpoint's oldestXID: 654
Latest checkpoint's oldestXID's DB: 16416
Latest checkpoint's oldestActiveXID: 92143770
Latest checkpoint's oldestMultiXid: 1
Latest checkpoint's oldestMulti's DB: 16416
Time of latest checkpoint: Sat 12 Mar 2016 01:56:19 PM PST

Brian

In response to

Responses

Browse pgsql-bugs by date

  From Date Subject
Next Message nummervet 2016-03-14 08:06:37 BUG #14019: Security label TAB causes unexpected EOF and client session termination
Previous Message Tom Lane 2016-03-12 16:03:49 Re: BUG #14017: strange syntax error