From: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com> |
---|---|
To: | Kyotaro HORIGUCHI <horiguchi(dot)kyotaro(at)lab(dot)ntt(dot)co(dot)jp> |
Cc: | masao(dot)fujii(at)gmail(dot)com, pgsql-hackers(at)postgresql(dot)org |
Subject: | Re: [BUG] Checkpointer on hot standby runs without looking checkpoint_segments |
Date: | 2012-04-17 09:06:05 |
Message-ID: | 4F8D327D.3010107@enterprisedb.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
On 17.04.2012 09:50, Kyotaro HORIGUCHI wrote:
> This is new version of the patch.
> I replaced GetStandbyFlushRecPtr with GetXLogReplayRecPtr to
> check progress of checkpoint following Fujii's sugestion.
The reason we haven't historically obeyed checkpoint_segments during
recovery is that it slows down the recovery unnecessarily if you're
restoring from a backup and you replay, say, one week's worth of WAL
files. If for example you have checkpoint_segments=10 and
checkpoint_timeout='15 mins' in the server that generated the WAL, you
would be constantly performing a restartpoint if you trigger one every
10 segments.
You could argue that you should obey checkpoint_segments in a standby
server that's caught up with the master, but AFAICS the patch doesn't
try to detect that.
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
From | Date | Subject | |
---|---|---|---|
Next Message | Pavel Stehule | 2012-04-17 09:50:55 | Re: Clobbered parameter names via DECLARE in PL/PgSQL |
Previous Message | Pavel Stehule | 2012-04-17 07:12:41 | Re: Why can't I use pgxs to build a plpgsql plugin? |