Re: Autovacuum to prevent wraparound tries to consume xid

Alexander Korotkov <a.korotkov@postgrespro.ru>

From: Alexander Korotkov <a.korotkov@postgrespro.ru>

To: pgsql-hackers <pgsql-hackers@postgresql.org>

Date: 2016-05-19T14:14:31Z

Lists: pgsql-hackers

Attachments

fix_vac_truncate_clog_xid_consume_2.patch (application/octet-stream) patch

On Mon, Mar 28, 2016 at 2:05 PM, Alexander Korotkov <
a.korotkov@postgrespro.ru> wrote:

> After some debugging I found that vac_truncate_clog consumes xid just to
> produce warning.  I wrote simple patch which replaces
> GetCurrentTransactionId() with ShmemVariableCache->nextXid.  That
> completely fixes this situation for me: ShmemVariableCache was successfully
> updated.
>

I found that direct reading of ShmemVariableCache->nextXid is not corrent,
it's better to use ReadNewTransactionId() then.  Fixed version of patch is
attached.

------
Alexander Korotkov
Postgres Professional: http://www.postgrespro.com
The Russian Postgres Company

Commits

Avoid repeating loads of frozen ID values.
- dd0183469bb7 17.0 landed
Close race condition between datfrozen and relfrozen updates.
- f222349c4ec0 12.19 landed
- 70cadfba0c70 13.15 landed
- 2ca19aa8165c 14.12 landed
- 92685c389d54 16.3 landed
- 7c5915c4b16c 15.7 landed
- f65ab862e3b8 17.0 landed
Fetch XIDs atomically during vac_truncate_clog().
- 2d2e40e3befd 9.6.0 cited