> We recently migrated most of those users from POP3 to IMAP. > Before, IMAP was barely used.
I have about the same number of users on one of the QmailToaster servers. We used to have most users on POP3, but have migrated most to IMAP-SSL. We noticed no load difference. > A reasonable thought, but if it were DNS it would be consistent. > There's a local caching nameserver on the mailserver. We're doing the same, this helped big time. I have a few questions: 1) Are you using the latest updates for your distsro? 2) Can you e-mail a pic of your stats-toaster of the IMAP graphs? 3) Can you paste a copy of `rpm -qa | grep toaster` 4) What does your memory usage look like when your users are using imap big time? 5) Does your hosts table have just localhost and the fqdn defined? 6) What kind of delay do you see when you telnet, does it take a while for the IMAP banners to pop up? 7) Have you migrated from an older version of QmailToaster (such as the old 1.0 branch w/ horde) 8) Are most of your users accessing the QmailToaster's IMAP through some sort of firewall? IE Users -> PIX DMZ -> QmailToaster 9) Does it make a difference if you disable the iptables firewall (temporarily for testing of course) 10) Is your resolv.conf pointing to localhost or the machines static ip for your caching name server? Sorry for the large number of questions. If you have some time, maybe we can speak on IRC later. Thanks, Erik --------------------------------------------------------------------- QmailToaster hosted by: VR Hosted <http://www.vr.org> --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]