| |||||||||||||||||||||||||||||||
|
Hi, We use a commercial anti-virus/spam filter configured as a content filter. Everything runs smooth and even though we are processing some 30 or 40 thousand emails, the CPU load is typically down in the level of the noise (i.e., hardly enough to measure). Now suddenly I'm seeing periods where the process load jumps way up, the five minute average sometimes hitting the mid 40's. When this happens, the Postfix mailq can sometimes get several thousand deep. I think the problem is the content filter, mainly because I can stop Postfix and the load only very slowly drops. There are no specific errors (at least none that generate log entries). The one thing I do notice is in the maillog entries when Postfix relays the mail to the filter. This is an example: relay=localhost.linfield.edu[127.0.0.1], delay=1780, status=sent (250 OK, sent 474CE652_12080_9_2 ) My question is what exactly is the delay value? What is it measuring? When everything is running nicely, the delay value seldom even gets as high as 10 but in the above example, it's 1780. Wjat exactly is it? Thanks, -- Rob Tanner UNIX Services Manager Linfield College, McMinnville OR
| ||||||||||||||||||||||||||||||
© 2004-2008 readlist.com