<html>
Hi,<br><br>
Initially when we started using MailScanner, we ran sendmail as,
"<b>/usr/lib/sendmail -q5m -oQ/mail/hosts/acc/smtp_out</b>" to
process the outgoing queue directory. But this was not speedy
enough for us. It took very long for mail to get
delivered. So we changed it to:<br>
"<b>/usr/lib/sendmail -OForkEachJob=True -OMaxDaemonChildren=20
-oQ/mail/hosts/acc/smtp_out -q1m</b>"<br>
This configuration for processing the outgoing queue is very effective in
clearing the queue.<br><br>
However, occasionally when our SMTP usage increases by 3 or 4 folds (over
150,000 SMTP transactions per day), the ForkEachJob configuration for
processing the sendmail outgoing queue, overwhelms our 4 processor CPU
(large box) and it runs out of processes. <br><br>
So, here are my questions:<br><br>
1. It is not clear to us as to what initiates the sendmail delivery
process in the outgoing queue. Once the mail has been scanned and
put in the outgoing queue, what signals sendmail to deliver the mail - is
it the MailScanner?<br><br>
2. For a beefy multi-processor box with a lot of power, is there a
recommended way to spread the load among the processors without running
the risk of "resource starvation".<br><br>
Thanks.<br><br>
Vasantha<br>
</html>
-------------------------- MailScanner list ----------------------<br>
To leave, send leave mailscanner to <a href="mailto:jiscmail@jiscmail.ac.uk">jiscmail@jiscmail.ac.uk</a><br>
Before posting, please see the Most Asked Questions at<br>
<a href="http://www.mailscanner.biz/maq/">http://www.mailscanner.biz/maq/</a> and the archives at<br>
<a href="http://www.jiscmail.ac.uk/lists/mailscanner.html">http://www.jiscmail.ac.uk/lists/mailscanner.html</a><br>