Results 1 to 4 of 4
Thread: failover server
-
10-26-2007, 03:09 AM #1Newbie
- Join Date
- Sep 2004
- Posts
- 14
failover server
I'd like to create a failover server, a complete and identical copy (as much as possible) of my primary server, which becomes "primary" when the first server dies.
I've been reading a lot about this subject and i've seen an already working method which uses rsync and MySQL replication.
Has anyone managed to create such a system? is there someone with experience or can point me to a relevant web site?
Thank you.
-
10-26-2007, 03:27 AM #2Web Hosting Master
- Join Date
- Oct 2004
- Location
- Kerala, India
- Posts
- 4,771
Have a check here.
http://dev.mysql.com/doc/refman/5.0/...ion-howto.htmlDavid | www.cliffsupport.com
Affordable Server Management Solutions sales AT cliffsupport DOT com
CliffWebManager | Access WHM from iPhone and Android
-
10-26-2007, 03:38 AM #3Newbie
- Join Date
- Sep 2004
- Posts
- 14
yeah well, that is the least of my problems, MySQL replication is easy, compared to the other problems that a failover server has to deal with.
here are some of the issues involved:
1) what happens when the primary server is back online? does it act as a backup server from now on? (means the two servers have identical resources)
2) what happens if a file is changed in both servers? which one do you keep?
3) what happens with emails?
4) do all of the above depend on how fast you make the switch?
5) is rsync capable of coping with the above issues?
-
10-26-2007, 12:05 PM #4Hosting Billing Master
- Join Date
- May 2003
- Location
- California, USA, Earth
- Posts
- 1,098
You want heartbeat, check out http://www.linux-ha.org/
1. In High Availability configurations, it's usually recommended that if the primary server comes back online the secondary stay in control of resources. This is set via the config files, it can work either way.
2. That's a big problem, and for HA configurations is why you should stick with active/passive configuration.
3. You configure heartbeat to use what are called "resources" sendmail/apache and whatever else would be off on the passive node, and active only on the active node. If there is a failure on the primary node, the secondary node would gain these resources automatically.
4. The switch is very fast, it's unlikely it would skip a ping.
5. rsync works ok for most things, syncing data from the primary node to the secondary node.. but you'll usually run this from a cron which means that it's possible that if the primary node goes down, the secondary won't have an exact copy of your data. DRBD is a solution for this.