Skipping queue run -- load average too high

[Martin Hepworth]

Harry Hanson wrote:
>>hmm seems to be page faulting quite a bit, which is unusual
>>for something relatively high CPU/memory wise.
>>
>>what checks have you got on SA and MS? Esp what checks are
>>running for RBL's and pyzor?
>
>
> During that time SA was disabled.
>
> Most MS settings are default, as I felt it best to ask for advice from this
> list before making any changes.
>
> RBL's are not being used via mailscanner; they are rejected by sendmail, so
> those that are listed don't even make it thru MA for processing.
>
>
>>sendmail is saying the machine is too busy to process mail,
>>hence the log messages..
>>
>>you say you've got the disk configured as RAID 0 (striping),
>>is this hardware or software RAID (vinum?).
>
>
> It is hardware raid.
>
>
>>have you turned on softupdates on the filesystem containing
>>the spool files - this can make alot of difference as regards I/O.
>
>
> I am unfamiliar with this.. How would I check?
>
>
>>How have you split up the filesystems? single / or /, /usr,
>>/home and /var?
>
>
> Yes, /var is on different physical drives.
>
>
>>Just wondering why you choose FreeBSD 5.1 as this is still
>>considered 'unstable'? The current 'stable' release is 4.9.
>>But my tests seem to indicate 5.x tree is much faster than 4.8...
>
>
> Honeslty, this was not my choice, but I figured since it was already there,
> and most of what I read indicated it's quite stable, I decided to stick with
> it and check it out.
>
>
>>I run FreeBSD 4.8 with Exim 4.24 and MS 4.24 (no RBL's/pyzor)
>>using Sophos-Savi and ClamAV with Mailwatch and the mysql DB
>>all on the same machine - celeron 600mhz, single ATA1-100
>>disk and a single / partition.
>>runs 9,000 messages a day without breaking above 1.5 on load average.
>
>
> load averages:  0.28,  0.27,  0.23
>
> Yes, this seems unusual, which is why I am deferring to the mailing list
> members hoping someone has some useful insight :)
>

OK check out tunefs to turn on softupdates for the /var filesystem
(you'll need to unmount it first. so you're prob going have to drop to
single user mode..)

Would be interesting to see what iostat and vmstat report when run at
the same time. Also getting a dump from 'top' and netstat might help to
see what's occurring at the time things go awry.

the default for sendmail to stop accepting mail is when the load average
goes about 12 (and to queue at load avg above 8)..so perhaps sendmail is
getting the load average wrong??/

from the sendmail.cf - default FreeBSD 5.1...(/etc/mail/sendmail.cf)

# load average at which we just queue messages
#O QueueLA=8


# load average at which we refuse connections
#O RefuseLA=12

I'd check that these are still correct..


--
Martin Hepworth
Snr Systems Administrator
Solid State Logic
Tel: +44 (0)1865 842300


**********************************************************************

This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.

This footnote confirms that this email message has been swept
for the presence of computer viruses and is believed to be clean.

**********************************************************************