= 80) then only
that session will use the looney sort_mem
It would be interesting to know if your machine is swapping.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 8: explain analyze
?
They are really prepared.
That depends on what version you are using. Older versions did what
Tom mentioned rather than sending PREPARE EXECUTE.
Not sure what version that changed in.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net
for a value?
there's been a thread on -hackers recently about checkpoint issues..
in a nut shell there isn't much to do. But I'd say give bizgres a
try if you're going to be continually loading huge amounts of data.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http
into some more hardware.. see if
you can borrow any or fabricate a poor man's equivalent for testing.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 5: don't forget to increase
real effect?
This doesn't allocate anything - it is a hint to the planner about
how much data it can assume is cached.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 9
. Instead you can do
something even better - compile PG normally and attach to it with
Shark (Comes with the CHUD tools) and check out its profile. Quite
slick actually :)
I'll keep people updated on my progress, but I just wanted to get
these issues out in the air.
--
Jeff Trout [EMAIL
gcc version 4.0.0 (Apple Computer, Inc. build 5026)
The snapshot on ftp.psotgresql.org (dated 8/29) also runs with no
optimization.
No cflags are set.
need to see anything from config.log?
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net
to
this machine, but be warned gprof on tiger is pretty useless from
what I've seen.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 6: explain analyze is your friend
(loadavg) with low cpu usage means you are IO bound.
Either change some queries around to generate less IO, or add more
disks.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 2: Don't
thing that may be a factor is the network - when doing
explain analyze it doesn't have to transfer the dataset to the client.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 2
datasets. It also tries to figure
out hte size you want (2x ram) - the original bonnie is limited to 2GB.
--
Jeff Trout [EMAIL PROTECTED]
http://www.jefftrout.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 6: explain analyze is your
On May 3, 2006, at 10:16 AM, Vivek Khera wrote:
On May 3, 2006, at 9:19 AM, Jeff Trout wrote:
Bonnie++ is able to use very large datasets. It also tries to
figure out hte size you want (2x ram) - the original bonnie is
limited to 2GB.
but you have to be careful building bonnie++ since
there that is
kicking the performance down the tube.
If it was PG's fault it wouldn't be stuck uninterruptable.
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 9: In versions below 8.0
to be going with hw r5, which went against what I
thought going in - read perf is more important for my usage than write.
I'm still not sure about that software 10 read number. something is
not right there...
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http
controller.
If you look in the pgsql-performance archives a week or two ago
you'll see a similar thread to
this one - in fact, it is also about a dl385 (but he had a 5i
controller)
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http://www.stuarthamm.net
to have data in the cache while seeing
how long it takes for the battery to drain :)
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP 3: Have you checked our extensive FAQ
on pgfoundry, it sort of simulates a pg index
scan, which is probably what you'll want to focus on more than seq
read speed.
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http://www.stuarthamm.net/
---(end of broadcast)---
TIP
, we need the random perf
much more than seq, but I can see the opposite being true in a
warehouse workload.
btw, the tool I wrote is here http://pgfoundry.org/projects/pgiosim/
--
Jeff Trout [EMAIL PROTECTED]
http://www.dellsmartexitin.com/
http://www.stuarthamm.net
are not yet in RAM?
Do you have sar or vmstat reports?
If you do not have good random io performance log replay is nearly unbearable.
(I've run into this before many times)
Also, what io scheduler are you using? if it is cfq change that to deadline or
noop.
that can make a huge difference.
--
Jeff
19 matches
Mail list logo
