On Jun 30, 2014, at 4:57 PM, Jeff Frost wrote:
>
> On Jun 30, 2014, at 4:04 PM, Tom Lane wrote:
>
>> Ah ... that's more like a number I can believe something would have
>> trouble coping with. Did you see a noticeable slowdown with this?
>> Now that we
On Jun 30, 2014, at 4:04 PM, Tom Lane wrote:
> Ah ... that's more like a number I can believe something would have
> trouble coping with. Did you see a noticeable slowdown with this?
> Now that we've seen that number, of course it's possible there was an
> even higher peak occurring when you sa
On Jun 30, 2014, at 1:46 PM, Jeff Frost wrote:
>> So it seems like we have a candidate explanation. I'm a bit surprised
>> that StandbyReleaseLocks would get this slow if there are only a dozen
>> AccessExclusiveLocks in place at any one time, though. Perhaps that
>&
On Jun 30, 2014, at 1:39 PM, Tom Lane wrote:
>
>
>> Another item of note is the system catalogs are quite bloated:
>> Would that cause the replica to spin on StandbyReleaseLocks?
>
> AFAIK, no. It's an unsurprising consequence of heavy use of short-lived
> temp tables though.
>
Yah, this h
On Jun 30, 2014, at 1:15 PM, Andres Freund wrote:
> On 2014-06-30 12:57:56 -0700, Jeff Frost wrote:
>>
>> On Jun 30, 2014, at 12:54 PM, Matheus de Oliveira
>> wrote:
>>
>>>
>>> On Mon, Jun 30, 2014 at 4:42 PM, Jeff Frost wrote:
>>>
On Jun 30, 2014, at 12:54 PM, Matheus de Oliveira
wrote:
>
> On Mon, Jun 30, 2014 at 4:42 PM, Jeff Frost wrote:
> And if you go fishing in pg_class for any of the oids, you don't find
> anything:
>
> That is probably because you are connected in the wrong database.
On Jun 30, 2014, at 12:32 PM, Tom Lane wrote:
> Jeff Frost writes:
>> Sampling pg_locks on the primary shows ~50 locks with ExclusiveLock mode:
>
>> mode | count
>> --+---
>> AccessExclusiveLock |11
>&
On Jun 30, 2014, at 12:17 PM, Jeff Frost wrote:
>>
>> already is quite helpful.
>>
>> What are you doing on that system? Is there anything requiring large
>> amounts of access exclusive locks on the primary? Possibly large amounts
>> of temporary relations
On Jun 30, 2014, at 11:39 AM, Andres Freund wrote:
> On 2014-06-30 11:34:52 -0700, Jeff Frost wrote:
>> On Jun 30, 2014, at 10:29 AM, Soni M wrote:
>
>>> It is
>>> 96.62% postgres [.] StandbyReleaseLocks
>>> as Jeff said. It runs quite
On Jun 30, 2014, at 10:29 AM, Soni M wrote:
>
>
>
> On Tue, Jul 1, 2014 at 12:14 AM, Andres Freund wrote:
>
> My guess it's a spinlock, probably xlogctl->info_lck via
> RecoveryInProgress(). Unfortunately inline assembler doesn't always seem
> to show up correctly in profiles...
>
> What wo
On Jun 30, 2014, at 9:14 AM, Heikki Linnakangas wrote:
> On 06/30/2014 05:46 PM, Soni M wrote:
>> Here's what 'perf top' said on streaming replica :
>>
>> Samples: 26K of event 'cpu-clock', Event count (approx.): 19781
>> 95.97% postgres [.] 0x002210f3
>
>
ernel.org/patch/825212/
and the bad behavior stopped. Best performance was with a 3.5 kernel with
the patch removed.
--
Jeff Frost
CTO, PostgreSQL Experts, Inc.
Phone: 1-888-PG-EXPRT x506
FAX: 415-762-5122
http://www.pgexperts.com/
--
Sent via pgsql-performance mailing list (pgsql-p
On Tue, 13 Jan 2009, Tom Lane wrote:
Jeff Frost writes:
On Tue, 13 Jan 2009, Tom Lane wrote:
It would change the size of the sample for the table, which might
improve the accuracy of the stats. IIRC you'd still get the same number
of histogram entries and most-common-values for the
On Tue, 13 Jan 2009, Tom Lane wrote:
Jeff Frost writes:
So, my question is, should changing the stats target on the shape column
affect the stats for the content_id and content_type columns?
It would change the size of the sample for the table, which might
improve the accuracy of the stats
columns? Also, why
does the index on content_id win out over the compound index on
(content_type, content_id)?
"index_blips_on_content_id" btree (content_id)
"index_blips_on_content_type_and_content_id" btree (content_type,
content_id)
--
Jeff Frost, Owner
Fr
On Fri, 31 Oct 2008, Gregory Stark wrote:
Tom Lane <[EMAIL PROTECTED]> writes:
Jeff Frost <[EMAIL PROTECTED]> writes:
Tom Lane wrote:
Huh. That does sound like it's a version-to-version difference.
There's nothing in the CVS log that seems related though. Are you
wil
Tom Lane wrote:
> Jeff Frost <[EMAIL PROTECTED]> writes:
>
>> On Thu, 30 Oct 2008, Tom Lane wrote:
>>
>>>> Any idea why I don't see it on 8.3.4?
>>>>
>>> I think it's more likely some small difference in your t
s that making the indexes before the updates seems to make the
planner happy!
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 916-647-6411 FAX: 916-405-4032
--
Sent via pgsql-performance mailing list (pgsql-performance@pos
Hmm ...
> I'm not sure if that's sufficient if there are other concurrent
> transactions; but it's certainly necessary.) Another possibility is
> to create the indexes just after data load, before you start updating
> the columns they're on.
>
>
Tha
Tom Lane wrote:
> Jeff Frost <[EMAIL PROTECTED]> writes:
>
>> I've run across a strange problem with PG 8.3.3 not using indexes on a
>> particular table after building the table during a transaction.
>>
>
> This may be a HOT side-effect ... is
3.4. I didn't see
any mention of a fix for this sort of thing in 8.3.4's release notes. I
was wondering if this is a known bug in 8.3.3 (and maybe other 8.3.x
versions) and just didn't make it into the release notes of 8.3.4?
--
Jeff Frost, Owner <
Tom Lane wrote:
Jeff Frost <[EMAIL PROTECTED]> writes:
I have two postgresql servers. One runs 8.3.1, the other 8.3.3. On the 8.3.1
machine, the index scans are being planned extremely low cost:
Index Scan using ix_email_entity_thread on email_entity (cost=0.00..4.59
right. I probably didn't mention that the slow one has been
analyzed several times. In fact, every time adjusted the statistics target
for that column I analyzed, thus the eventually better, but still inaccurate
estimates toward the bottom of the post.
--
Jeff Frost, Owner <[EMAIL
0 on the server with the 4.59 cost
estimate.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
--
Sent via pgsql-performance mailing list (pgsql-performance@postgresql.org)
To make changes to you
On Wed, 14 Nov 2007, Merlin Moncure wrote:
On Nov 14, 2007 5:24 PM, Alan Hodgson <[EMAIL PROTECTED]> wrote:
On Tuesday 13 November 2007, Jeff Frost <[EMAIL PROTECTED]> wrote:
Ok, Areca ARC1261ML. Note that results were similar for an 8 drive RAID6
vs 8 drive RAID10, but I don&
On Wed, 14 Nov 2007, Alan Hodgson wrote:
On Tuesday 13 November 2007, Jeff Frost <[EMAIL PROTECTED]> wrote:
Ok, Areca ARC1261ML. Note that results were similar for an 8 drive RAID6
vs 8 drive RAID10, but I don't have those bonnie results any longer.
Version 1.03 -
--Random Create
-Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
files /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP /sec %CP
16 6655 16 + +++ 5755 12 7259 17 + +++ 5550 12
--
Jeff Frost, Owner <[EMAIL
On Sat, 6 Oct 2007, Tom Lane wrote:
"Jeff Frost" <[EMAIL PROTECTED]> writes:
Before analyze it seems to choose Bitmap Heap Scan on episodes
current_episode, but after it chooses Index Scan Backward using
index_episodes_on_publish_on on episodes current_episode.
Have you t
On Fri, 5 Oct 2007, Stephen Frost wrote:
* Jeff Frost ([EMAIL PROTECTED]) wrote:
Here are the plans:
It's probably just me but, honestly, I find it terribly frustrating to
try and read a line-wrapped explain-analyze output... I realize it
might not be something you can control in
1 loops=6229)
-> Bitmap Heap Scan on episodes current_episode
(cost=2.34..15.65 rows=11 width=8) (actual time=0.007
..0.016 rows=13 loops=6229)
Recheck Cond: (season_id = $0)
Filter: ((publish_on IS NOT NU
r friend
---(end of broadcast)---
TIP 3: Have you checked our extensive FAQ?
http://www.postgresql.org/docs/faq
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-
if you're not
doing this with regularity and strongly consider enabling autovacuum.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
eport against the log and post us the explain analyze from your
slow queries.
And if Ron is indeed local, it might be worthwhile to contact him. Someone
onsite would likely get this taken care of much faster than we can on the
mailing list.
--
Jeff Frost, Owner <[EMAI
DB on every web page, you might consider playing
with the commit_delay and commit_siblings parameters in the postgresql.conf.
Also, if you're doing multiple inserts as separate transactions, you should
consider batching them up in one transaction.
--
Jeff Frost,
unning multiple postmasters on the
same machine that can speak to the postgresql.conf knobs more specifically.
I'd still suggest you upgrade to at least 8.1.8.
Thanks
-Original Message-
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Jeff Frost
Sent: Thursday, Apri
ove:
http://archives.postgresql.org/pgsql-performance/2007-03/msg00104.php
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
TIP 5:
pg_xlog+OS. Your
workload may vary, but it's definitely worth testing. The system in question
had 1GB BBU.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---
ntly going from SQL_ASCII to UTF8.
In 8.1 you can do this:
SELECT datname,
pg_size_pretty(pg_database_size(datname)) AS size
FROM pg_database;
In 7.4, you'll need to install the dbsize contrib module to get the same info.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Fr
On Thu, 1 Mar 2007, Alex Deucher wrote:
On 3/1/07, Jeff Frost <[EMAIL PROTECTED]> wrote:
On Thu, 1 Mar 2007, Alex Deucher wrote:
>> >> Postgresql might be choosing a bad plan because your
>> effective_cache_size
>> >> is
>> >> way off (it'
t and costs quite a bit less than a SAN.
Is the SAN being shared between the database servers and other servers? Maybe
it was just random timing that gave you the poor write performance on the old
server which might be also yielding occassional poor performance on the new
one.
--
Jeff Fro
e analyze running merrily
along in the background. It's probably not as bad off as you think. At least
this query isn't 10x. :-)
Run these again for us after analyze is complete.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC h
On Thu, 1 Mar 2007, Alex Deucher wrote:
On 3/1/07, Jeff Frost <[EMAIL PROTECTED]> wrote:
On Thu, 1 Mar 2007, Alex Deucher wrote:
>> Vacuum? Analayze? default_statistics_target? How many shared_buffers?
>> effective_cache_size? work_mem?
>>
>
> I'm runnin
nie tests? Probably want to tune
random_page_cost as well if it's also at the default.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadca
On Thu, 1 Mar 2007, Alex Deucher wrote:
On 3/1/07, Jeff Frost <[EMAIL PROTECTED]> wrote:
On Thu, 1 Mar 2007, Joshua D. Drake wrote:
> Alex Deucher wrote:
>> Hello,
>>
>> I have noticed a strange performance regression and I'm at a loss as
>> to wha
one of the
10x slower queries would probably be handy.
What do you mean by "created from scratch rather than copying over the old
one"? How did you put the data in? Did you run analyze after loading it?
Is autovacuum enabled and if so, what are the thresholds?
--
Jeff Frost, Owne
ze output from 7.4.5 and 8.2.3 for the
query in question?
Also, is the hardware the same between 7.4.5 and 8.2.3? If not, what is the
difference?
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-
. In our case, the battery backed write cache seemed to
remove the need for a separate WAL disk, but someone elses workload might
still benefit from it.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 65
faster than
ext3, but of course you could likely go with another filesystem yet and be
even slightly faster as well. :-)
I guess the real moral of the story is that you can probably use one big ext3
with the default config and it won't matter much more than 1-2% if you have a
BBU.
--
J
On Tue, 9 Jan 2007, Jim C. Nasby wrote:
On Thu, Dec 28, 2006 at 02:15:31PM -0800, Jeff Frost wrote:
When benchmarking various options for a new PG server at one of my clients,
I tried ext2 and ext3 (data=writeback) for the WAL and it appeared to be
fastest to have ext2 for the WAL. The
lways faster than the other two
options. I didn't test any other filesystems in this go around.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(
On Sat, 16 Dec 2006, Arjen van der Meijden wrote:
On 16-12-2006 4:24 Jeff Frost wrote:
We can add more RAM and drives for testing purposes. Can someone suggest
what benchmarks with what settings would be desirable to see how this
system performs. I don't believe I've seen an
benchmarks with what settings would be desirable to see how this system
performs. I don't believe I've seen any postgres benchmarks done on a quad
xeon yet.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7
only pairs
after that.
A valid question. Does the caching raid controller negate the desire to
separate pg_xlog from PGDATA?
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-64
rest.
I could only find the 6 disk RAID5 numbers in the archives that were run with
bonnie++1.03. Have you run the RAID10 tests since? Did you settle on 6 disk
RAID5 or 2xRAID1 + 4XRAID10?
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsult
d how much BBU cache can
you put in it? Oh, does it use the good ole megaraid_mbox driver as well?
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(
d DB which is mostly read
intensive, but occassionally has large burts of write activity due to new user
signups generated by the marketing engine.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650
laces?
Currently, I'm looking at Penguin, HP and Sun (though Sun's store isn't
working for me at the moment). Maybe I just need to order a Penguin and then
buy the controller separately, but was hoping to get support from a single
entity.
--
Jeff Frost, Owner <[
would tune postgres to avoid using the
CPU.
Neil
On 12/07/06, Jeff Frost <[EMAIL PROTECTED]> wrote:
On Wed, 12 Jul 2006, Neil Hepworth wrote:
> I am using version PostgreSQL 7.3.10 (RPM:
> postgresql73-rhel21-7.3.10-2). Unfortunately vacuumdb -a -v does not
> give the FSM
ew version ?
- Original Message -----
From: "Jeff Frost" <[EMAIL PROTECTED]>
To: "Neil Hepworth" <[EMAIL PROTECTED]>
Cc:
Sent: Wednesday, July 12, 2006 10:27 AM
Subject: Re: [PERFORM] High CPU Usage - PostgreSQL 7.3
On Wed, 12 Jul 2006, Neil Hepworth wrote:
On Tue, 11 Jul 2006, Jeff Frost wrote:
On Wed, 12 Jul 2006, Neil Hepworth wrote:
You might also want to turn on autovacuum and see if that helps.
What's your disk subsystem like? In fact, what's the entire DB server
hardware like?
By the way, how big does the temp table get
AND start <
TO_TIMESTAMP((TO_TIMESTAMP('2006-07-12 11:02:13.865444+1000',
'-MM-DD 00:00:00.0')::timestamp - INTERVAL '10080 MINUTE'),
'-MM-DD 00:00:00.0')::timestamp
MAIN LOOP TOTAL deleteExpiredData: 505142
MAIN LOOP TOTAL generateStatistics: 515611
ot;.eppairdefnid = "inner".eppairdefnid)
-> Seq Scan on ftone (cost=0.00..23583.33 rows=1286333 width=10)
(actual time=0.04..2299.94 rows=1286333 loops=1)
-> Hash (cost=0.00..0.00 rows=1 width=4) (actual
time=206.01..206.01 rows=0 loops=1)
-> Seq Scan on fttemp
d thousand rows it still takes tens of minutes
with high CPU. My database does have a lot of tables (can be several
thousand), can that cause performance issues?
Thanks,
Neil
---(end of broadcast)---
TIP 3: Have you checked our extensive FAQ?
ast few lines of VACUUM VERBOSE say?
Also, are you running ANALYZE with the vacuums or just running VACUUM? You
still need to run ANALYZE to update the planner statistics, otherwise
things might slowly grind to a halt. Also, you should probably consider
setting up autovacuum and upgrading
8.1 for better performance overall.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
TIP 6: explain analyze is your friend
07346
(1 row)
(231907346-231894522)/300 = 42.74666666 TPS
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)
y to continue the restore process from where it left off?
Thanks,
Brendan Duddridge | CTO | 403-277-5591 x24 | [EMAIL PROTECTED]
ClickSpace Interactive Inc.
Suite L100, 239 - 10th Ave. SE
Calgary, AB T2G 0V9
http://www.clickspace.com
On Apr 20, 2006, at 3:19 PM, Jeff Frost
rate, it's
going to take about 10 hours to restore our database.
Most of the time, the server is not using very much CPU time or I/O time. So
I'm wondering what can be done to speed up the process?
Brendan,
Where are the WAL files being stored and how are they being read back?
--
If you respond off-list, I'll summarize and post the results back.
Thanks for any input.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
/lmbench
As numbers from lmdd are seen on this frequently.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
TIP 3: Ha
bit user-space; data warehouse type tests
(data >> memory); and web prefs test (active data RAM)
What specific benchmarks should be run, and what other things should be
tested? Where should I go for assistance on tuning each database, evaluating
the benchmark results, and re-tuning them?&qu
ft Lockspam to fight spam, and you?
http://www.polesoft.com/refer.html
---(end of broadcast)---
TIP 2: Don't 'kill -9' the postmaster
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsu
What's the current status of how much faster the Opteron is compared to the
Xeons? I know the Opterons used to be close to 2x faster, but is that still
the case? I understand much work has been done to reduce the contect
switching storms on the Xeon architecture, is this correct?
--
http://www.powerpostgresql.com/PerfList/
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Phone: 650-780-7908 FAX: 650-649-1954
---(end of broadcast)---
TIP 1: if posting/read
spx?i=2163&p=2
It's a little old, as it's listing an Opteron 150 vs 3.6 Xeon, but it does
show that the opteron comes in almost twice as fast as the Xeon doing
Postgres.
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostconsultingllc.com/
Ph
On Tue, 19 Apr 2005, J. Andrew Rogers wrote:
I don't know about 2.5x faster (perhaps on specific types of loads), but the
reason Opterons rock for database applications is their insanely good memory
bandwidth and latency that scales much better than the Xeon. Opterons also
have a ccNUMA-esque I
e's a way to
extract that sort of info from other metrics it keeps in the stats table?
Maybe a script which polls the stats table and correlates the info with stats
about the system in /proc?
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consulting, LLC http://www.frostc
stems, though I have heard from a
few Postgres folks that a dual Opteron is 2.5x as fast as a dual Xeon. I
would think that AMD would be all over that press if they could show it, so
what am I missing? Is it a bus speed thing? Better south bridge on the
boards?
--
Jeff Frost, Owner <[
xRAID5 for data
c) 1xRAID10 for OS/xlong/data
d) 1xRAID1 for OS, 1xRAID10 for data
e) .
I was initially leaning towards b, but after talking to Josh a bit, I suspect
that with only 4 disks the raid5 might be a performance detriment vs 3 raid 1s
or some sort of split raid10 setup.
--
Jeff Fr
s the performance of
gigE good enough to allow postgres to perform under load with an NFS mounted
DATA dir? Are there other problems I haven't thought about? Any input would
be greatly appreciated.
Thanks!
--
Jeff Frost, Owner <[EMAIL PROTECTED]>
Frost Consul
80 matches
Mail list logo