Start a Conversation

Unsolved

This post is more than 5 years old

2042

February 11th, 2016 08:00

RecoverPoint with XtremIO and VMware SRM

Hello,

What options do we have to achieve application-consistent PITs when using native snap based replication with XtremIO? Protected servers are Microsoft SQL Server 2012, virtualized on vSphere 5.5. No RDMs being used. The entire DR is managed by VMware SRM. Please share any tech note or whitepaper discussing this scenario.

28 Posts

February 11th, 2016 14:00

Rich,

How do we quiesce the database? If you suggest KVSS utility, I would like to know if this method is still relevant to a virtualized SQL Server  2012, RP 4.1.2, and XtermIO 4.0.1.

2 Intern

 • 

1.1K Posts

February 11th, 2016 14:00

The process is no different than with any other array or replication mode.

Quiesce the SQL database, initiate the bookmark (this in turn initiates a new snapshot), wait for it to complete and then resume the database.

Regards,

Rich Forshaw

Consultant Corporate Systems Engineer - RecoverPoint & VPLEX (EMEA)

Data Protection and Availability Solutions

EMC Europe Limited

Mobile: 44 (0) 7730 781169 44%20(0)%207730%20781169>

E-mail: richard.forshaw@emc.com

Twitter: @rw4shaw

2 Intern

 • 

522 Posts

February 11th, 2016 16:00

You could leverage the KVSS utilities as well to perform the bookmark in your script - I have done this with RDM's before and the process should be the same depending on how you have your VMDK's laid out on the Datastores. You would then need to use VSI PiT recovery with the SRM integration to chose that app-consistent image during failover/mount on the target side.

2 Intern

 • 

1.1K Posts

February 12th, 2016 08:00

I'm with Keith on this re the provision of a application-consistent image with RDM devices and using the VSI plugin for RP to select the image in conjunction with SRM.

Where VMFS is being used I cannot advocate using vss unless the VM is running Windows and the vdisk to VMFS datastore to LUN is a 1-2-1 mapping. You also need to ensure that the SCSI target ID's are the same.

2 Intern

 • 

522 Posts

February 12th, 2016 08:00

yah - I have done it with the 1-2-1 mapping of the VMDK's to datastores like Rich said and that was what I meant with regards to the VMDK layout on the DB's.....I was just typing on my phone and shortened it..haha. I can provide some test documentation that I did with RDM's and the KVSS utility if that helps...if you meed the VMDK layout you should be able to apply it in the same manner.

No Events found!

Top