SRM 5.5 using EMC SRDF. Recovery completes but reprotect fails

[RyanH84]

Hi all,

I'm fairly new to SRM but have been plugging away setting up a new environment which is using:

vCenter 5.5 u1  (Both Sites)

Solutions Enabler 8.0.1.0-1953

EMC SRDF Adapater (SRA) = 5.6.0  VMAX 20k

EMC SRDF SRA Utilities (VSI) = 5.7.0.2

SRM = 5.5.1.8569

I've been following documentation provided by Cody Hosterman  ~which is superb and has gotten me almost to the end.  I've been able to successfully fail over (Recovery) a VM using a test Recovery Plan from Site A to Site B. It worked great.

We are not using TimeFinder (not licensed) and so when we perform a recovery the expected behavior happened that the SRDF link has been broken. The state of the R1 device at Site A has become Write Disabled. The R2 device in Site B has become Read/Write and the state of the device group is in "Failed Over". This is in Asynchronous SRDF mode.

Upon trying to "Reprotect" the group, I see errors relating to configuring the storage to reverse direction.

http://i.imgur.com/zWiGtT5.png

I've checked the EmcSrdfSra log as suggested on the now primary site SRM server and it appears that there is an error shortly after a new consistency group is created.

http://i.imgur.com/5hZT8HQ.png

I've raised a support call with EMC with this directly as a first port of call but it is in it's infancy so I thought I'd ask if anyone has any clue what might be going wrong here. I've exhausted my googleFU skills and am at the stage of trying to get some experienced eyes on the issue. I'm happy to provide more information if needed or answer any questions anyone might have.

Really hope someone can assist!

Thanks!

Ryan

[RyanH84 · Accepted solution]

Hi,

I had some progress with this so I thought I would come back and share what it took to fix this issue.

Essentially, in our environment we have a VMAX in each site. We now have two array manager servers with Solutions Enabler (one has Unisphere web) installed. The reason for this is that we have a VCE Vblock, and VCE have provided an array manger server. In addition, we have built our own for the purposes of SRM and other such things.

In essence, we had not configured the new Solutions Enablers servers fully and were missing GNS on them. We were also setting up SRDF pairs and most importantly: DEVICE GROUPS on the old array managers manually. These device groups were not replicating to the new SE servers, but the other config (Volumes, etc) were! Therefore SRM was trying to talk to our new boxes and only really getting half the information about the volumes and most importantly the device groups.

As soon as we properly configured the device groups on each site for the Volumes that were being used for SRDF replication, everything started to work. Since then we have been back in and re-checked the configuration on the new array manager servers and everything is good. This resolved the problems I have posted above.

Hope this helps someone in the future!

Ryan