SMVI 2.0 - fails when performing a restore of vmfs volume

[daniel_kaiser]

Hi,

Hopefully someone can shed some light on the issue we're having.  I'm able to use SMVI to backup NFS and VMFS datatstores/VM machines, but when I do a restore of a VM machine on a VMFS datastore it fails.  No restore issues with VM machines that are hosted on NFS, just VMFS.

We're running the following in our environment

OnTap 7.3.2

SnapManager for Virtual Infrastructure 2.0 (installed on vCenter server)

vSphere 4.0 w/update 1

Messages in the Task pane of SMVI

One or more mount requests did not succeed.  Please check vSphere Client for any error messages.

Failure in DatastoreMountAction:  One or more mount requests did not succeed.  Please check vSphere for any error messages.

Operation transitioning to abort due to prior failure.

Operation aborted.

Operation failed:  One or more mount requests did not succeed.  Please check vSphere Client for any error messages.

Messages in the Recent Tasks of vCenter Client

Rescan all HBA

Rescan VMFS

I see these two actions in the vCenter client when i submit the restore job.

Any help would greatly be appreciated.  Thanks.

[daniel_kaiser]

I noticed that a new Igroup was created called ESXhostname_smvidg_fcp and i assume this was created during the failed restore process.  The igroup settings appear to be identical to the Igroup we created for ESX.

[wodgersmith]

Did you get anywhere with this?

I am having the same issue. The backup seems to work fine and I can see the snapshot tasks appearing in the Virtual Center recent tasks but I see nothing when I try and mount the SNAP for restore.

I get a message that the Mount Request failed and to check the VSphere Client logs, however there is nothing reaching the VSphere client??

[amritad]

Hi

Is this an NFS datastore or a VMFS datastore that you're trying to mount? An NFS datastore mount operation would require FlexClone to be licensed.

REgards

Amrita

[eric_barlier]

Daniel,

what licenses have you got on this controller? I think NFS and SMVI requires a flexclone license. I know we had an issue where NTAP had to

issue us with a license to make SMVI and NFS work. Im pretty sure its flexclone but not 100% sure.

Eric

[ogra]

Hi,

Also check if you running VMFS over iSCSI the connectivity.

You might not able to reach the ESX server to mount the backup for restoration.

To isolate the Problem, simply try mounting the backup directly.

Cheers !!!

[amritad]

Hi

Flexclone licesnse is required for:

1) NFS datastore mounts

2)NFS VMDK restore to an alternate datastore

3)Single file restore

Backups and restores will not fail for other scenarios.

Regards

Amrita

[daniel_kaiser]

All,

We're running the controllers in an active/active environment and have flexclone licensed on both controllers.  The VMFS volumes are over Fibre Channel and no iSCSi is used in our environment.  Could the issue be somewhere on the fibre channel/switch configuration?  Thanks again for all your suggestions and please keep it coming.

[keitha]

Which ESX host are you selecting when you do the VM restore? Does that ESX server have access to the same FC LUN as the VM was backed up from? What will happen is SMVI will FlexClone the backup then connect the FlexCone to the ESX server you give it, then copy the VM from the FlexClone to the source VMFS. Thus it needs access to both the FlexClone and the original VMFS LUN.

Keith

[daniel_kaiser]

In the 'ESX Host ID' field, I'm selecting the same esx server that I backed up in the backup job.

[amritad · Accepted solution]

Hi

Is ALUA enabled? Also are you able to see the datastore name properly in the inventory tab?

Regards

Amrita

[daniel_kaiser]

Yes we have ALUA enabled and I can see the datastore that i want backed up in the datastore.

[amritad]

Hi,

So there is a releae note for ALUA enabled LUNs. so this isn't applicable right?

SnapManager for Virtual Infrastructure sometimes cannot

see datastores from LUNs mapped to ALUA-enabled

initiator groups

When you click Backup or Inventory in the SnapManager for Virtual Infrastructure user interface,

some datastores might be missing. These datastores are from LUNs mapped to ALUA-enabled

initiator groups.

Workaround:

Rescan all HBAs from each ESX host and then reopen the SnapManager for Virtual

Infrastructure user interface.

Regards

Amrita

[daniel_kaiser]

My specific issue has been resolved with the installation of SMVI 2.0P1D1.

http://now.netapp.com/NOW/download/software/snapmanager_vi/2.0P1D1/

SnapManager for Virtual Infrastructure - 2.0P1D1

Date Posted: 06-APR-2010

SnapManager for Virtual Infrastructure 2.0P1D1
04/05/2010

Bugs fixed in this release:

2.0P1D1
=======

388092:  High CPU utilization


405310: Null Pointer Exception when SMVI cannot properly track a vCenter task


387359:  Failure in MountCloneSetupTaskWrapper


403263 : Retry VMware quiesced snap, then force non-quiesced snapshot if quiesced VMware quiesced snapshot fails


2.0P1
=======

383747 - HostLunMapper should check the transport type instead of just casting it and hoping


380467 - SMVI restore operation causes duplicate entrys in /etc/exports


390010 - ApplyRetention deletes recent snapshot if expected snapshot is missing


392529 - If vCenter returns a null reference for a request to get a VM's UUID, SMVI does not          check the response appropriately


393240 - VmSnapshotDeleteAction: java.lang.reflect.UndeclaredThrowableException


375255 - SMVI unable to mount a datastore; mount operation fails!


379489 - SQLException is thrown and causes SMVI operation to fail

[oosterbaan]

We seem to have the exact same problems, but installing the P1D1 version generates a 'invalid service name' error about some openSCmanager service and a few other error messages.

did you install anything prior to this p1d1 version?