SRM and EMC Clarion

[clarges]

Hi, quick question, I am setting up an SRM enviroment using 2 emc cx-420, i have configured the storage on both sites and configured the zones etc, I have presented and configured the luns on the production site as vmfs volumes

my question is, do you need to add the luns in as vmfs on dr site before configuring srm or is there an intelligent process on the san replication adapter to do this

[depping · Accepted solution]

You will need to create a VMFS volume on the protected site, this volume gets replicated to the recovery site by utilizing the arrays replication technology. When you do a failover the luns will be published to the recovery site, hba's will be rescanned and the volume will be resignatured. I would suggest start reading the SRM admin pdf which contains a wealth of information:

http://www.vmware.com/pdf/srm_10_admin.pdf

Duncan

Blogging: http://www.yellow-bricks.com

If you find this information useful, please award points for "correct" or "helpful".

[clarges]

cheers duncan, been through the documentation a good bit, just needed a little clarification, vmfs created already on protection site and was not sure if it needed to be done on DR site as well or just raw disks presented to esx

[bladeraptor]

Hi

SRM and VMware site protection is a funny one.

People perhaps rightly presume that as virtualization changes paradigms in som many other parts of an infrastructure that it should do the same with site recovery.

SRM is a precursor to a wider range of VMware provided frameworks that allow the vendors to deliver functionality to meet some specific VMware / wide industry outcome. The vStorage initiative is an example of this

However in the SRM case as it is the vendors driving the mechanics of recovery if you have ever done this in the real world the principles are the same here with SRM.

So you have a target site where volumes are mounted to and formatted by the production hosts.

At the array level those volumes are replicated to a remote site where they can (in the case of SRM) be mounted as effectively 'hidden' 'read only' volumes on the DR hosts.

Upon SRM DR invocation the DR side othe replication relationship is promoted and the read only DR copy has its flags stripped and a refresh of the ESX hosts reveals the volumes to the DR environment.

On the old production side if it is still live, the volumes are stripped from the ESX environment and placed in a 'hidden' 'read only' capacity on the old production hosts.

Full failover is fairly drastic so to advance the idea of regular testing, snapshots of the source and target volumes can be used to present the production data to the remote site in a non-disruptive fashion where the 'dummy' VM images can then be tested

Very similar to mounting a backup job to a backup proxy - just no read-only :]

So if you have played with this technology and understand the concepts in the physical world chances are you will understand the basic mechanics that the cleverness of SRM is masking - then again this is just V1!

Regards

Alex Tanner