From: | Jeff Trout <threshar(at)torgo(dot)978(dot)org> |
---|---|
To: | Jeffrey W(dot)Baker <jwbaker(at)acm(dot)org> |
Cc: | Steve Poe <spoe(at)sfnet(dot)cc>, paul(at)oxton(dot)com, pgsql-performance(at)postgresql(dot)org |
Subject: | Re: [SPAM?] Re: PG8 Tuning |
Date: | 2005-08-12 12:18:27 |
Message-ID: | B4099474-30C8-4549-A289-2EC4FD1FF4CB@torgo.978.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-performance |
On Aug 11, 2005, at 12:58 PM, Jeffrey W. Baker wrote:
> Like Mr. Stone said earlier, this is pure dogma. In my experience,
> xlogs on the same volume with data is much faster if both are on
> battery-backed write-back RAID controller memory. Moving from this
> situation to xlogs on a single normal disk is going to be much
> slower in
> most cases.
>
This does also point one important point about performance. Which is
a touch unfortunate (and expensive to test): Your milage may vary on
any of these improvements. Some people have 0 problems and
incredible performance with say, 1000 shared_bufs and the WAL on the
same disk.. Others need 10k shared bufs and wal split over a 900
spindle raid with data spread across 18 SAN's...
Unfortunately there is no one true way :(
The best bet (which is great if you can): Try out various settings..
if you still run into problems look into some more hardware.. see if
you can borrow any or fabricate a "poor man"'s equivalent for testing.
--
Jeff Trout <jeff(at)jefftrout(dot)com>
http://www.jefftrout.com/
http://www.stuarthamm.net/
From | Date | Subject | |
---|---|---|---|
Next Message | Jeffrey Tenny | 2005-08-12 22:37:45 | Re: Mostly read performance (2 replies) |
Previous Message | Steve Poe | 2005-08-12 08:47:08 | [SPAM?] Re: PG8 Tuning |