Re: BUG #9161: wal_writer_delay is limited to 10s

From: Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
To: Clemens Eisserer <linuxhippy(at)gmail(dot)com>
Cc: pgsql-bugs(at)postgresql(dot)org
Subject: Re: BUG #9161: wal_writer_delay is limited to 10s
Date: 2014-02-18 19:20:20
Message-ID: 9250.1392751220@sss.pgh.pa.us
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-bugs

Clemens Eisserer <linuxhippy(at)gmail(dot)com> writes:
> Thanks a lot for the in-depth information.
> Is the 8KB boundary configureable at compile time?

There is a configure option for XLOG_BLCKSZ, but it won't go beyond 64K
which is probably not enough to fix your problem. (I'm not sure offhand
if there are technical reasons for 64K in particular as the limit, but
in any case I suspect making it really large would create performance
issues of its own.)

I think what you'd really want is to tweak the walwriter logic so that
it avoids writing individual blocks and instead waits for wal_buffers
to be, say, half full before writing.

regards, tom lane

In response to

Browse pgsql-bugs by date

  From Date Subject
Next Message Joshua Yanovski 2014-02-18 19:25:54 Re: BUG #9227: Error on SELECT ROW OVERLAPS ROW with single ROW argument
Previous Message Tom Lane 2014-02-18 19:15:34 Re: BUG #9227: Error on SELECT ROW OVERLAPS ROW with single ROW argument