Mailscanner and server load

[kfliong]

<x-flowed>
I have just implemented SURBL. Server load shot up and stays around 30
making websites unaccessable. I have now edited SURBL to only check on
SPAMCOP_URI_RBL. I have disabled the other 3. Now my load averages around
12. This is still quite high.

BTW, after some monitoring using "top", I notice that my system is quite
RAM intensive. But "top" can't tell for sure. What other tools can I use to
see whether the highload is due to lots of disk accesses (due to not having
enough RAM).

At 03:26 PM 26/8/2004, you wrote:
>On Thu, 2004-08-26 at 03:55, David J. Duffner - NWCWEB.com wrote:
> >         #3 - DON'T PANIC on load averages.  We're using MRTG and
> MailScanner
> > MRTG and they don't agree with load averages, so you're not seeing a
> > realtime accurate report!
><snip>
> >     But overall you're seeing
> > an illusion, we finally got a grip on it as MailScanner MRTG samples
> > at a different rate and is giving us a REAL average to work with on
> > loads.
>
>Just to explain that a little.  MailScanner-MRTG samples every five
>minutes and uses the five minute average load - thats the second load
>average figure given by top and w.  This is handy as it means it takes
>the average over its reporting period so is less prone to being messed
>up by brief spikes.  A word of caution - older MSMRTG's didn't work like
>that (it used the one minute average - which tended to be misleading).
>
>Upping the RAM to 1G is a good idea, depending on your message
>throughput you may be able then to put the MailScanner work directory in
>tmpfs - that gives a significant performance boost so long as you have
>enough physical ram (if it causes it to start swapping thats a bad
>thing).
>You need to get a feel for whether your system is CPU bound or IO bound,
>the above trick will really help with disk IO related issues, network IO
>issues can be helped by running a caching nameserver on the box itself
>(very easy, theres an rpm for caching-nameserver I think, then you just
>point resolv.conf and 127.0.0.1).
>A load average of 3 isn't necessarily a problem anyway, so long as mail
>is flowing with acceptable latency and you can cope with spikes.  Given
>that your machine was able to battle through the backlog that built up
>during your upgrade (hence the 30 load av) I'd say its not too
>concerning (I'd still add ram, do the tmpfs thing and run a
>caching-nameserver though).
>
>Kevin
>
>
>
>
>BMRB International
>http://www.bmrb.co.uk
>+44 (0)20 8566 5000
>_________________________________________________________________
>This message (and any attachment) is intended only for the
>recipient and may contain confidential and/or privileged
>material.  If you have received this in error, please contact the
>sender and delete this message immediately.  Disclosure, copying
>or other action taken in respect of this email or in
>reliance on it is prohibited.  BMRB International Limited
>accepts no liability in relation to any personal emails, or
>content of any email which does not directly relate to our
>business.
>
>------------------------ MailScanner list ------------------------
>To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
>'leave mailscanner' in the body of the email.
>Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
>the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at jiscmail.ac.uk with the words:
'leave mailscanner' in the body of the email.
Before posting, read the MAQ (http://www.mailscanner.biz/maq/) and
the archives (http://www.jiscmail.ac.uk/lists/mailscanner.html).
</x-flowed>