Current Setup: server 1: perfect wheezy, apache2: ispc3 cp, dns, web, db. 8 core, 2 TB raid 1 server 2: perfect wheezy, apache2, courier: web, db, webmail (sogo), email. 8 core, 2 TB raid 1, with 1 TB in email data. backup is via back-res.sh on both servers, incremental daily, monthly full. You can imagine the time it takes to backup server 2 on the monthly. I'm trying to figure out the best way to make this setup more fault tolerant at very least for th email server and for backups. - setup a secondary server for each (master./master, master/slave...etc)? - what's the best way to keep a secondary email server (remember it's not dovecot)....rsync? and only run full backups on the secondary server? - Could I get by with a single 8 core server and virtualize 2 x 4 cores to be secondary servers to server 1 and 2? As you can see, I'm curious how others have production environments setup, and maybe that will help me with proper questions.
Personally I would recommend to switch to dovecot. I migrated all my old courier server to dovecot in the past years. With Dovecot you can use the internal dsync feature to replicate the emails. If you have to stay with yourier, then use either rsync for one way sync or the setup from the ispconfig cluster tutorial.
Thank you for the guidance Till. Switching to dovecot has been on my to do list for the past year but didn't want to break something that was already working in production. Any chance there is an updated how to switch guide? I know you built a script over a year ago, I just couldn't find a step by step.
