Here is a snapshot of my environment:
- SQL Server 2005 Standard
- Windows Server 2003 Enterprise R2
- VMWare ESX 3.0 (2 node cluster)
- EqualLogic PS iScsi SAN
- Unitrend DPU
My company is implementing a new ERP system using SQL 2005 Standard as it's DB. The topic of high availability/DR came up in our discussion and really created a lot more questions than answers. Since we use Standard Edition I made a suggestion to use the new Mirroring capability that comes with it to mirror to one of our remote offices in Iowa. Realizing that this would require us to purchase and build a new server, license the OS and SQL, and get a bigger pipe connecting our sites (p-to-p T-1 currently) a co-worker suggested that we do something different. His suggestion is to rely on VMWare snapshots and/or EqualLogic snapshots to bring the server back on line and use the SQL backups to restore the DB to production state. I can see the merit in his solution as it is much more cost-effective, but I can't help but question the idea of relying solely on snapshots to get us back up and running. Doing initial discovery for our SLA the owners of the ERP system said that the acceptable downtime is no more than 3 hours. Obviously this all depends on how good our backups/snapshots are so I just want to kind of get a feel for what anyone else thinks. Thanks.
- SQL Server 2005 Standard
- Windows Server 2003 Enterprise R2
- VMWare ESX 3.0 (2 node cluster)
- EqualLogic PS iScsi SAN
- Unitrend DPU
My company is implementing a new ERP system using SQL 2005 Standard as it's DB. The topic of high availability/DR came up in our discussion and really created a lot more questions than answers. Since we use Standard Edition I made a suggestion to use the new Mirroring capability that comes with it to mirror to one of our remote offices in Iowa. Realizing that this would require us to purchase and build a new server, license the OS and SQL, and get a bigger pipe connecting our sites (p-to-p T-1 currently) a co-worker suggested that we do something different. His suggestion is to rely on VMWare snapshots and/or EqualLogic snapshots to bring the server back on line and use the SQL backups to restore the DB to production state. I can see the merit in his solution as it is much more cost-effective, but I can't help but question the idea of relying solely on snapshots to get us back up and running. Doing initial discovery for our SLA the owners of the ERP system said that the acceptable downtime is no more than 3 hours. Obviously this all depends on how good our backups/snapshots are so I just want to kind of get a feel for what anyone else thinks. Thanks.