re: [rdiff-backup-users] Failover

[Greg Freemyer]

 >>  I'm new to rdiff-backup and thinks it's one of the most useful tools 
 >>  I've ever run across.

 >>  My question is:

 >>  Is anyone using this for a true failover mirror?  Is it possible to use 
 >>  this routine to push an entire configuration to a secondary box, then 
 >>  use the secondary box in case primary fails?

 >>  If anyone is doing this, can you give me an idea of your routine.

 >>  Thanks again

 >>  Mark Bulson
 >>  Charleston, SC

I'm not doing this rdiff-backup, but I do a fair amount of disaster recovery 
planning / HA clustering.

Assuming Linux:

For a LAN, drdb and heartbeat (Linux-HA) seem like a good pairing that can 
handle moderate data volumes. 

For a WAN, I would think rsync would work better than rdiff-backup.  You could 
use cron to sync up every hour or so.  

A big problem for the WAN env. is getting the IP routing to go to the right 
location upon a failover.  Linux-HA handles that in the LAN env.

====
Of course, if you have "Enterprise" class needs, the typical data redundancy 
solutions are:

    LAN: Shared SCSI or FC arrays   (IBM also makes a ServeRAID product which 
fits between these and drdb in price).

   WAN: Data Replication implemented via FC fabric and managed in the FC array. 
 

HTH
Greg
-- 
Greg Freemyer