SugarCRM Disaster Recovery strategies
I'm trying to figure out how to provide Disaster Recovery/Business Continuity for my Sugar Open Source installation.
As a first cut, I was thinking of a warm backup approach using hosting service A to handle the primary instance, and then send periodic filesystem and MySQL backups over to hosting service B and script things in such a way that the backup site on B can be brought up on demand, reflecting the latest backup.
This would give me the ability to offer a fallback, read-only site where people can consult the system data but would have to hold their updates aside until the primary system on hosting service A was back up.
A second cut was to allow primary status to flip-flop back and forth between A and B, at the cost of some extra scripting complexity.
Then I remembered my experience running some other software inside VMWare virtual machines. Maybe I should just take a performance hit, run Sugar inside a Linux VM, periodically snapshot the entire VM (RAM+disk+running SugarCRM app) and FTP the snapshots to a backup virtual hosting provider. Theoretically, the application can come right back up as of the last snapshot courtesy of VMWare with little reconfiguration (save perhaps some IP address jiggling). No MySQL database sync, no need to quiesce the primary SugarCRM app (the snapshots can be taken off a running VM!), no separate filesystem backup restore issues.
Does anybody out there have experience running Sugar and MySQL inside VMs? Are there any gotchas I need to watch out for?
If they work as advertised, the Disaster Recovery advantages will make VM hosting practically obligatory for Sugar because CRM is definitely a mission-critical piece of the average business.
John Benson
Codeasaurus Rex
Bookmarks