> So, I am looking into possibly using SquirrelMail to provide web mail > access for our website members it is currently running on a Dell 2400 > Power Edge dual processor, scsi, 60 GIG HD, 1.5 Gig RAM using White Box > Linux 3.0 ES for my OS. I have read the FAQ's for enhancing SQM > performance which I intend > to use. I will be using the latest Postfix with MYSQL for my MTA with > spamassasin > handling spam including clamav for virus filtering all in one box due to > budget limitations.
In my opinion, given that you've only got PIII's in that box, you've got too much for it to handle. It can certainly handle the mail (provided that 60GB disk is actually some form of RAID), but I've found the squirrelmail/IMAP end to be more cpu-intensive than one might expect. (But note that you wouldn't be able to run Exchange on that config either.) In our configuration, we've got a Dual PIII 1000+2GB 2450 as the MTA/IMAP/POP3 box for ~4,000 users and 3xPE1550's with single PIII 1000's+1GB load balanced for SQM with about 1000 daily SQM users and each web server serving around 50k hits daily. We're not using any PHP optimizers yet. On the IMAP end, things can slow down to the point where the web application has a noticeable performance hit as well, but that's rare. We deliver around 50,000 emails/day to these 4,000 accounts. Some notes: - With 4,000 users depending on one application to access their email, you must have hardware redundancy. Load balance several lower-end web servers (with LVS or something) rather than having one beefy box. - With 4,000 users, depending on how much they use email (probably a lot!), make sure you've got a good I/O path to your storage. U160 SCSI at a minimum, some sort of striping/mirroring preferable. We've got a 14-disk JBOD hooked up to a PERC4/DC and it performs nicely, but on our former array, a 4-disk RAID-5 over U2W on a PERC2/DC, performance was unacceptable. Avoid RAID5 if possible. - Use a PHP optimizer on your frontend. - Keep the IMAP service and SQM on different hardware if at all possible: both can be cpu-intensive depending on the options you enable. - Note that our 3 node LVS cluster serves 1,000 users nicely and you're talking about 4,000 users. Sorry to be the bearer of bad news here -- I just don't think you have the hardware to do it. For such a number of users, the end-to-end design has to be scalable and resilient. John -- John Madden UNIX Systems Engineer Ivy Tech State College [EMAIL PROTECTED] ------------------------------------------------------- This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170 Project Admins to receive an Apple iPod Mini FREE for your judgement on who ports your project to Linux PPC the best. Sponsored by IBM. Deadline: Sept. 13. Go here: http://sf.net/ppc_contest.php -- squirrelmail-users mailing list Posting Guidelines: http://squirrelmail.org/wiki/wiki.php?MailingListPostingGuidelines List Address: [EMAIL PROTECTED] List Archives: http://news.gmane.org/thread.php?group=gmane.mail.squirrelmail.user List Archives: http://sourceforge.net/mailarchive/forum.php?forum_id)95 List Info: https://lists.sourceforge.net/lists/listinfo/squirrelmail-users
