VMware Cloud Community
cbhtechsupport
Contributor
Contributor

Presenting replicated LUN's to DR site. Rescan failure.

Hey guys,

Thought I would post this here to see if anyone has seen or experienced this issue we had during a DR test.

Background on environment:

2x sites (Head Office and DR) connected via Fibre

2x VMware vCenter Servers (One each site)

2x 6x Node ESXi 5.5 Clusters (One each site)

2x IBM V7000 SAN replicated async using Global Mirror

We use our DR site as a DEV/Test Environment outside of DR.

So this is what happened. On Saturday we simulated a DR at our Head Office site. By this I mean we shut down all VM's running on the cluster including the Head office vCenter server. All that was left running was the hosts. We then switch the direction of our mirroring making Head Office LUN's Read Only and making DR LUN's R/W. We the present these LUN's to the DR cluster.

So at this point we shutdown all DEV/TEST workload and there are only a handful of non DEV/TEST servers running on the DR cluster (DR vCenter, Domain Controller, Core network Services)

We then proceed to rescan the storage of a DR Host and add the DR LUN's one by one without resignaturing the LUN. (We did this once in the past when in ESX v4). This process of rescanning and adding the storage to the DR host took forever. And what I expected was once scanned and added to the first host the others do their normal rescan and automatically add the volumes. This didn't happen. It was taking in excess of 20mins per lun to scan in the add storage wizard.

On these DR hosts we have probably 80 LUN's assigned with the DR/DEV/TEST LUN's assigned together. Some are RDM's.

We ended up rolling back, switching the mirrors back and undoing everything.

We have done this process in the past when on vSphere v4.1 and expected it to work as planned but now I'm stumped.

Any ideas?

Cheers,

Garrick

Reply
0 Kudos
4 Replies
SG1234
Enthusiast
Enthusiast

interesting issue -- can you share the vmkernel logs from the hosts on the DR site?

Reply
0 Kudos
vNEX
Expert
Expert

Hello,

when you perform a manual DR you must first resignature the volume and next you have to reregister all VMs in respect of the new volume name and UUID.

All that work is for SMR ... it automatically handles resignature process as well as .vmx reference corrections ...when I see your site specs and LUN numbers it's definitely worth giving it a try.

Regards,

P.

_________________________________________________________________________________________ If you found this or any other answer helpful, please consider to award points. (use Correct or Helpful buttons) Regards, P.
Reply
0 Kudos
vNEX
Expert
Expert

Hi,

any updates on this issue?

Thanks,

P.

_________________________________________________________________________________________ If you found this or any other answer helpful, please consider to award points. (use Correct or Helpful buttons) Regards, P.
Reply
0 Kudos
cbhtechsupport
Contributor
Contributor

Ended up getting them mapped. We first set all our RDM's luns to perenially reserved. This sped up the rescan. The we mapped the luns to the farm of hosts and added them via esx-cli setting the resignaturing to false.

I think it may have been having trouble with the RDM's to be honest.

Reply
0 Kudos