On Fri, Dec 24, 2004 at 05:54:49AM -0500, Darryl Clarke wrote: > As for a 'switch' I would do this: > (assuming your servers aren't configured via dhcp and you have control > over the IPs) > On the secondary box have a cron that 'pings' the to see if its alive. > This process should take into consideration the time it takes for your > primary server to reboot -- including potential fsck's. ex. if it's > down for 1 minute don't switch, but if its down for > 20 minutes > switch over. If in the event the primary is non responsive the > secondary box can ifup eth0:1 -- an aliased (non automagic) interface > using the primary server's ip address. This way it will basically > takeover any incoming requests for the now non-responsive main server.
Interesting idea, but the two servers are not just in different networks, but different states. Both have static IP addresses. I'll stick to the "switch the DNS entry" method. Also do it manually -- with four people authorized to edit that entry, we can get to it pretty fast, and this is a noncommercial site, so an hour of downtime will rarely be a big deal. Thanks to all for your suggestions. I'll investigate the various packages. Oh, and I should have said: no need to copy me on responses, I do read debian-user. -- Carl Fink [EMAIL PROTECTED] Jabootu's Minister of Proofreading http://www.jabootu.com -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
