|
List Info
Thread: slowness
|
|
| slowness |

|
2006-06-14 22:27:59 |
Good luck
Let us know how you get on
Wayne
On Wed, 2006-06-14 at 16:20 -0600, Greg Woods wrote:
> On Wed, 2006-06-14 at 22:54 +0100, Wayne Lee wrote:
> > what about telnet to 127.0.0.1 25, does postfix
accept new mail?
>
> Yes. Also not a surprise since the active queue will at
times grow quite
> large.
>
> However, I may have found a pointer to the problem. I
have recipient
> address verification turned on, and I never realized
this worked by
> sending pseudo-mail messages (called "probe
messages"). These messages
> are also going through amavisd, which is not necessary.
I need to play
> with the address verification configuration parameters
to see if I can
> get the probe messages to use the downstream SMTP
server instead of
> going through amavisd. I don't know if this alone will
solve the problem
> or not, but it might, because there are a large number
of probe messages
> (recognizable because they have from=<postmaster ucar.edu>) in the
> active queue when delays are occurring.
>
> --Greg
>
>
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
|
| slowness |

|
2006-06-14 23:04:50 |
On Wed, 2006-06-14 at 23:27 +0100, Wayne Lee wrote:
> Let us know how you get on
Back to the drawing board. I got the probe messages to route
through the
downstream MTA, but the delays are still happening.
It's got to have something to do with the database since
the web
interface slows down at the same time. Any chance it could
be related to
having MyISAM tables instead of InnoDB?
--Greg
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
|
| slowness |

|
2006-06-14 23:28:43 |
Greg Woods wrote:
> It's got to have something to do with the database
since the web
> interface slows down at the same time. Any chance it
could be related to
> having MyISAM tables instead of InnoDB?
Highly likely, yes. MyISAM locks the whole table to do a
write
operation, whereas InnoDB only needs to lock a specific row,
so clearly
you're going to experience more contention-related delays
with MyISAM
tables on a busy system.
--
Robert LeBlanc <rjl renaissoft.com>
Renaissoft, Inc.
Maia Mailguard <http://www.maiamail
guard.com/>
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
|
| slowness |

|
2006-06-15 00:59:27 |
|
The most contention problems we've seen have been with the bayes db
when bayes is stored in the file system. MySQL hasn't been a problem
for us, but you can take a look at mysqlard for a general idea how sql
is perfoming. You should turn up amavis logging for a while to see
which steps are causing the slowdown.
-Blake
Robert LeBlanc wrote:
renaissoft.com" type="cite">
Greg Woods wrote:
It's got to have something to do with the database since the web
interface slows down at the same time. Any chance it could be related to
having MyISAM tables instead of InnoDB?
Highly likely, yes. MyISAM locks the whole table to do a write
operation, whereas InnoDB only needs to lock a specific row, so clearly
you're going to experience more contention-related delays with MyISAM
tables on a busy system.
_______________________________________________
Maia-users mailing list
renaissoft.com">Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
| slowness |

|
2006-06-14 23:44:52 |
On Wed, 2006-06-14 at 16:28 -0700, Robert LeBlanc wrote:
> MyISAM locks the whole table to do a write
> operation, whereas InnoDB only needs to lock a specific
row,
Wow, thank you, that is great to know. That means there is
reason to
believe that converting to InnoDB might help.
--Greg
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
|
| slowness |

|
2006-06-15 12:45:09 |
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Greg Woods wrote:
> On Wed, 2006-06-14 at 16:28 -0700, Robert LeBlanc
wrote:
>> MyISAM locks the whole table to do a write
>> operation, whereas InnoDB only needs to lock a
specific row,
>
> Wow, thank you, that is great to know. That means there
is reason to
> believe that converting to InnoDB might help.
I keep coming back to that timing you showed where it seems
stuck in the queue,
and once amavisd-maia gets it, it zips right through. I
don't see how the
database is involved here... unless amavisd-maia doesn't
report anything until
it has stuck the message into the database...
At any rate, it would still be a good idea to convert to
InnoDB.
- --
David Morton
Maia Mailguard - http://www.maiamailguard
.com
Morton Software Design and Consulting - http://www.dgrmm.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
iD8DBQFEkVZVUy30ODPkzl0RAnokAKDSm9Ru8ynbGv/D12iH7x7oHwZklACg
mR5w
HqehLXj6yN/oV34bhilN/JQ=
=qi8q
-----END PGP SIGNATURE-----
_______________________________________________
Maia-users mailing list
Maia-users renaissoft.com
http://www.renaissoft.com/mailman/listinfo/maia-users
|
|
[1-6]
|
|
|
about | contact Other archives ( Real Estate discussion Medical topics )
|