RecoverPoint VMFS replicated Datastores are slow to mount

[VMDogg]

We are currently in the process of doing Disater Recovery testing and we are using RecoverPoint to replicate the LUNs (datastores) to our DR sites SAN.  Anyway, when we cut over to our DR site and mount the LUNs and go to add the datastores, it takes a considerble amount of time (10-15 minutes)for vSphere to see them.

Anyway, I believe that the long scan times are because the LUNs that are being presented already have a VMFS volume signature on it.  Because this is a “snapshot copy” of the production system and it has all the VMFS volume header info from the production system, it must be taking an extra-long time to scan.

Anyway, would changing any of the vCenter Storage filter settings help? Please see page 116 on the ESX Configuration Guide:    http://www.vmware.com/pdf/vsphere4/r40/vsp_40_esxi_server_config.pdf

Any suggestions for speeding up this process would be greatly appreciated.

Thanks,

ba

[kjb007]

If they eventually are showing up on their own, then snapshot/resignature settings will not help.  If that were the case, then the datastores would not come back at all, and would require you to force mount to make them available.

Are you seeing any errors in your logs from when you present the LUNs, rescan the HBA's, and when the datastores show?  Make sure there is no existing reservation on these LUNs.

-KjB

[JoaoPereira]

Hi VMDogg

I´ve the same problem after DR process with asyncronous replication.

Have you some idea about this problem?

JP

[JoaoPereira]

Hi VMDogg,

Did you try rescan directly in vmhba# at ESX host?

esxcfg-rescan vmhba#

or

via PowerCLI in the same machine that you´ve vCenter installed, if you have a cluster:

PowerCLI C:\Program Files\VMware\Infrastructure\vSphere PowerCLI> Get-Cluster -name "cluster" | Get-VMHost | Get-VMHostStorage -RescanAllHBA -rescanVMFS -refresh

JP