Hello -
We have several environments all running SQL Server 2008 on top of ESXi 5.1. The end-goal is to replicate an exact copy of all the SQL databases from one environment to a 2nd environment. We don't necessarily want to replicate the entire server image, as there are differences between the environments that need to remain after the refresh.
We currently use an in-house developed process that uses Windows and T-SQL scripting to rip compressed backups of the SQL DBs, write them to an archive server, then restore to the 2nd environment on a scheduled basis (some DBs daily, some DBs weekly). We're currently on a 1GBe network, and the process obviously all happens over the wire from the Prod environment to the archive server, then from the archive server to the Pre-Prod environment.
the process works Ok for now, but we're currently "refreshing" about 2TB of data daily, and the process takes about 1 hr 40 mins. The application is growing quickly, and I anticipate the size will double in the next six months. Since the speed of the process is mostly network-bound, I'm concerned in the next year we'll hit a point where the data refresh will span several hours and become impractical do on a daily basis.
So my question is, is it possible to do the following with Orchestrator:
- Selectively snapshot a VMs VMDKs on the source virtual servers
- Copy the snapshot to another local VMFS volume (on shared-storage between the two hosts)
- Mount the copied VMDKs on the destination virtual servers
- Kick off a script (fired off by Orchestrator) on the destination virtual servers to complete the process by mounting the newly attached disks in Windows, then mounting the MDFs/LDFs in SQL Server
Have you ever heard of something like that being done with Orchestrator and VMDKs? The obvious advantage is keeping all the data movement on the SAN instead of compressing pushing over the network, downloading, and decompressing.
Thanks,
- Jeff