On 18 Feb 2009, at 12:48 am, Ruslan Zakirov wrote:
First time I see something like that. May be it's better to ask on
mysql lists. And sure you should keep digging into a reason. My
practice shows that there is no such thing as suddenly in IT. There
is always some change that turn world up
On Tue, Feb 17, 2009 at 05:52:23PM -0500, Jesse Vincent wrote:
An excellent question. I have a sneaking suspicion there's some DB
corruption in there somewhere :/ But maybe it's just a database bug.
After chasing some red herrings we found something interesting:
mysql select count(*) from
On Wed, Feb 18, 2009 at 06:13:08PM +, Tim Cutts wrote:
On 18 Feb 2009, at 5:54 pm, Dave Holland wrote:
On Tue, Feb 17, 2009 at 05:52:23PM -0500, Jesse Vincent wrote:
An excellent question. I have a sneaking suspicion there's some DB
corruption in there somewhere :/ But maybe it's
We migrated our RT server to RT 3.8.2 a week ago, and all went well
until today, when suddenly we've been hit by a massive performance
problem. It seems that almost any query which touches the Attachments
table takes an absolute age to complete. For example, I've just
logged into the
On Tue 17.Feb'09 at 16:51:57 +, Tim Cutts wrote:
We migrated our RT server to RT 3.8.2 a week ago, and all went well
until today,
From what version?
when suddenly we've been hit by a massive performance
problem. It seems that almost any query which touches the Attachments
On Tue 17.Feb'09 at 17:47:43 +, Dave Holland wrote:
Tim's just got into his car to go home, but I can provide some
answers...
On Tue, Feb 17, 2009 at 12:29:10PM -0500, Jesse Vincent wrote:
On Tue 17.Feb'09 at 16:51:57 +, Tim Cutts wrote:
We migrated our RT server to RT 3.8.2 a
On Tue, Feb 17, 2009 at 12:59:17PM -0500, Jesse Vincent wrote:
And just to be sure, you ran the commands in UPGRADING.mysql?
I believe so (Tim did it).
# Query_time: 240 Lock_time: 0 Rows_sent: 4 Rows_examined: 27
SELECT DISTINCT main.Id AS id, main.Filename AS filename,
I believe so (Tim did it).
# Query_time: 240 Lock_time: 0 Rows_sent: 4 Rows_examined: 27
That just looks...crazy. Can you try an OPTIMIZE TABLE Attachments; ?
(It may take ... a while.)
pgpUvNTAP2WCd.pgp
Description: PGP signature
___
Tim's just got into his car to go home, but I can provide some
answers...
On Tue, Feb 17, 2009 at 12:29:10PM -0500, Jesse Vincent wrote:
On Tue 17.Feb'09 at 16:51:57 +, Tim Cutts wrote:
We migrated our RT server to RT 3.8.2 a week ago, and all went well
until today,
From what
On Tue, Feb 17, 2009 at 8:47 PM, Dave Holland d...@sanger.ac.uk wrote:
Tim's just got into his car to go home, but I can provide some
answers...
On Tue, Feb 17, 2009 at 12:29:10PM -0500, Jesse Vincent wrote:
On Tue 17.Feb'09 at 16:51:57 +, Tim Cutts wrote:
We migrated our RT server to
On 17 Feb 2009, at 6:09 pm, Dave Holland wrote:
On Tue, Feb 17, 2009 at 12:59:17PM -0500, Jesse Vincent wrote:
And just to be sure, you ran the commands in UPGRADING.mysql?
I believe so (Tim did it).
Indeed. All done pukka according to the upgrade docs.
Tim
--
The Wellcome Trust
On 17 Feb 2009, at 6:16 pm, Jesse Vincent wrote:
I believe so (Tim did it).
# Query_time: 240 Lock_time: 0 Rows_sent: 4 Rows_examined: 27
That just looks...crazy. Can you try an OPTIMIZE TABLE Attachments; ?
(It may take ... a while.)
Might as well do that now, since we're out of
On 17 Feb 2009, at 6:16 pm, Jesse Vincent wrote:
I believe so (Tim did it).
# Query_time: 240 Lock_time: 0 Rows_sent: 4 Rows_examined: 27
That just looks...crazy. Can you try an OPTIMIZE TABLE Attachments; ?
(It may take ... a while.)
OK, OPTIMIZE TABLE done, but performance on
Now, I may be being silly here, but how can a table with only 806180
id's in its primary key have more than 3 million rows? And counting
them takes three minutes, which really isn't many rows a second...
Is this stored in InnoDB by chance? If it were me, I would mysqldump all
databases,
OK, OPTIMIZE TABLE done, but performance on SELECT COUNT(id) FROM
Attachments is still terrible:
This does sort of smell like an optimiser bug - it says it's using
indices, but the level of I/O being done by the machine suggests it's
actually performing a full table scan.
Indeed.
On 17 Feb 2009, at 6:39 pm, Ruslan Zakirov wrote:
I hope you see that mysql examined only 27 rows and spend 240 seconds
to do that. I suspect that it's related to very high IO activity that
were going parallel with this query and blocked mysql's IO requests.
That was my worry. I've shuffled
On 17 Feb 2009, at 10:52 pm, Jesse Vincent wrote:
OK, OPTIMIZE TABLE done, but performance on SELECT COUNT(id) FROM
Attachments is still terrible:
This does sort of smell like an optimiser bug - it says it's using
indices, but the level of I/O being done by the machine suggests it's
First time I see something like that. May be it's better to ask on
mysql lists. And sure you should keep digging into a reason. My
practice shows that there is no such thing as suddenly in IT. There
is always some change that turn world up side down.
optimize/analyze on InnoDB don't do the same
On 18 Feb 2009, at 12:48 am, Ruslan Zakirov wrote:
First time I see something like that. May be it's better to ask on
mysql lists. And sure you should keep digging into a reason. My
practice shows that there is no such thing as suddenly in IT. There
is always some change that turn world up
19 matches
Mail list logo