Also, look at the switches and/or routers your box is talking to. I spent time once replacing hardware inside a mailserver (which somewhat helped) but was still not happy with my overall performance.
I isolated the issue to a switch which was not dead but was having a lot of trouble; once I took that switch out of the equation, my insane load dropped dramatically. If you're getting a lot of retransmission anyplace, you can wind up with qpopper idling for very long periods of time before finally erroring out. -Peter -----Original Message----- From: Doug Hardie [mailto:[EMAIL PROTECTED] Sent: Wednesday, May 21, 2008 2:30 PM To: Sasa Cc: Subscribers of Qpopper Subject: Re: System resource On May 20, 2008, at 05:05, Sasa wrote: > Hi, I have a critical problem with qpopper 4.0.9, when qpopper service > is running on mail server the queue increase and I have more 200 mails > in spool !! I think that qpopper daemon use a too much system resource > (memory and CPU), if qpopper daemon is stopped in a few minutes the > queue is empty. > When qpopper daemon is running I have: > > [EMAIL PROTECTED] ~]# ps -ax|grep popper > Warning: bad syntax, perhaps a bogus '-'? See /usr/share/doc/ > procps-3.2.7/FAQ > 5026 ? S 0:00 /usr/local/sbin/popper -s -R -T600 > 5276 ? R 0:00 /usr/local/sbin/popper -s -R -T600 > 5277 ? S 0:00 /usr/local/sbin/popper -s -R -T600 > > > ..how can I solve the problem ?? Some more thoughts. I suspect the problem you are encountering can be caused by thrashing or a disk failing. You might want to use top and vmstat (whatever its equivalent is for your system) to check for excess swapping or disk I/O. Also check the system log for disk errors. Remapping and retries on drives are quite slow and really bottle up the system.
