- I am now only logging SMTP with the "Debug Messages" option checked.
- Diskeeper Defrag is running daily at 2am. I ran it manually this afternoon just to make sure it was working right. When it was finished the disk had zero fragmented files. 10 minutes later, the 120 user list received a message and the processor went to 100% for 2 minutes. - The mailing list is going to eternal users - I ran the pathping string as suggested on several IPs that were causing 10054 errors. They all had 90%+ packet loss towards the end of the tracert. Nothing remotely close to my network. Nothing that I can control. - Just to make sure this was not a DNS problem, we set up DNS services on a new server with nothing else on it. Pointed Imail to the new DNS IP. Same problem. - If we change the "number of recipients per message" to 25 instead of 5, we have 5 SMTP processes hogging all the processor instead of 20 SMTP processes hogging all the processor. If we change the number to 50, we have 2 SMTP processes each using 40% processor for between 1 and 2 minutes. - We are using the internal Imail database Thanks, Ted ----- Original Message ----- From: "Len Conrad" <[EMAIL PROTECTED]> To: <[EMAIL PROTECTED]> Sent: Friday, March 14, 2003 12:12 PM Subject: Re: [IMail Forum] Imail processor problems - continued > > > - SCSI Hard Drives. No partitioning. Defrag by Diskeeper. > > as we have said here many times, partitions: > > ** OS + apps > > ** dedicated swap > > ** dedicated spool/ > > ** mailbox storage > > ** and if you were a heavy webmail, a partition for webmail workspace > > Turn off all logging except SMTP (you gave stats for POP, so you must be > logging POP, turn it off). Turn on non-STMP logging for debugging. > > If we decide you're suffering from fragmentation, then think about a > Partition Magic to setup the partitions to what we think here is "best > practice" > > > - Mailboxes are using 4GB of total disk space. only 120 users take > >advantage of our "masochistic generosity" and use more than 10MB. > > ok, but if they are on that 50-post/day mailing list and heavy mail users, > then there will be lot of access to their mbx's in a very short time. > > > - The spool contains between 500 and 800 files on any given day, with no > >files older than 1 week. > > not too bad > > what about your defragging practice? > > With everything dumped in on partition, fragmentation could be horrendous. > > > - When processor utilization goes to 100%, the primary culprit is SMTPD > > receiving mail, and applying rules, etc > > >(10% to 40%) and between 10 and 20 SMTP processes (each between 0% and 20%). > >The 100% processor is most often triggered by a 120 user mailing list that > >receives about 50 posts per day. > > ok, that matches up, delivering to 120 users is a big dollop, but really, > not very much for your machine. > > and the SMTP would be delivering to the 120 list recipients. If a bunch of > those were local, then that would be a big shot of local disk access to > deliver to the mbx files. and if the partition was horribly fragmented.... > > > - can you give more explicate instructions on the looping batch file. > > :START > > pathping ..... > > GOTO START > > :)) > > help goto > > Len > > > To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html > List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ > Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/ > To Unsubscribe: http://www.ipswitch.com/support/mailing-lists.html List Archive: http://www.mail-archive.com/imail_forum%40list.ipswitch.com/ Knowledge Base/FAQ: http://www.ipswitch.com/support/IMail/
