Re: XLOG_PARAMETER_CHANGE handling of wal_log_hints

From: Heikki Linnakangas <hlinnakangas(at)vmware(dot)com>
To: Petr Jelinek <petr(at)2ndquadrant(dot)com>, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
Cc: PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: XLOG_PARAMETER_CHANGE handling of wal_log_hints
Date: 2015-01-15 18:57:36
Message-ID: 54B80DA0.9000704@vmware.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 01/07/2015 11:53 AM, Petr Jelinek wrote:
> On 07/01/15 00:59, Michael Paquier wrote:
>> On Wed, Jan 7, 2015 at 4:24 AM, Petr Jelinek <petr(at)2ndquadrant(dot)com> wrote:
>>> Hi,
>>>
>>> when I was fixing how commit_ts handles the XLOG_PARAMETER_CHANGE I noticed
>>> that for wal_log_hints we assign the value in ControFile to current value
>>> instead of value that comes from WAL.
>>>
>>> ISTM it has just information value so it should not have any practical
>>> impact, but it looks like a bug anyway so here is simple patch to change
>>> that.
>> Right. That's something that should get fixed in 9.4 as well..
>>
>> ControlFile->track_commit_timestamp = track_commit_timestamp;
>> The new value of track_commit_timestamp should be taken as well from
>> xlrec on master btw.
>
> That's part of the larger fix for CommitTs that I sent separately in
> response to the bug report from Fujii - there is much more to be done
> there for CommitTs than just this.

Pushed this part now. I'll let you and Fujii handle that larger fix. Thanks!

- Heikki

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Andres Freund 2015-01-15 19:15:42 Re: hung backends stuck in spinlock heavy endless loop
Previous Message Merlin Moncure 2015-01-15 18:53:24 Re: hung backends stuck in spinlock heavy endless loop