Snapvx for MS-SQL VM's

Hi Marvin,

Here are a couple docs that may help:

https://www.emc.com/collateral/white-papers/h14273-vmax3-timefinder-snapvx-microsoft-sql-server.pdf

https://www.emc.com/collateral/white-paper/h13576-ms-sql-vmax3-wp.pdf

I really don't have any SQL knowledge though so can't answer any specific questions. Found the docs on Google

Mike

Yes, I have seen these.  Unfortunately they do not address the specifics of using Snapvx on the volume that contains the SQL VMDK and any potential issues with regards to I/O in flight when the snap is executed, or the validity of the VMDK (OS and SQL) when a snap is restored.  I thought I once read that SQL might acknowledge a write prior to Vmware actually committing it to the array, which might cause a problem with the validity of the snap.

We are currently using Snapvx on a large MS-SQL DB on physical servers and have already tested restores of snaps.  It works fine.  Now we're inserting the VMware layer which is what I'm concerned about.

Thanks for your reply.

OK, got it. SQL is definitely not my area of expertise. I'll try to see if I can find someone who may be able to help out.

Mike

Hi Marvin,

  I have developed the paper https://www.emc.com/collateral/white-papers/h14273-vmax3-timefinder-snapvx-microsoft-sql-server.pdf. And as you correctly pointed out that focuses on Physical environments.

  But I can help you with your query regarding ESXi Cluster based deployments. Snap VX is application agnostic and works at storage group level so SnapVX on SQL Server VMDK based deployments would work similar to SnapVX on Physical storage group.

  Can you please send me an e-mail outlining your environment so that I can address your questions appropriately?

Thanks,

Udgith

Hi Udgith,

I would need to know the capacity consumption of the SNAPVX link target volume when the source device is VMDK-Thick Eager Zero.

Thanks,

Kesav