RDM LUNs not fully removed from VM

[GOgden]

I have an issue where RDMs disconnected from the Guest OS (in this case, Windows 2008R2 64bit) using SnapDrive 6.3.1 are sometimes not fully removed from the virtual machine. The disk is not visible in the OS or in SnapDrive so appears to have been removed, but going to Edit Settings on the VM will still show the RDM as connected, and the LUN still exists and appears mapped on the filer. I've also noticed when viewing the storage paths on the host a number of 'dead' paths, which I assume are old RDM connections.

This happens on both ESXi4 and ESXi5 hosts.

This can cause problems if a LUN is removed from the filer, as it appears to be no longer connected to the Guest OS, when the VM still thinks it is connected.

[kjb007]

Sounds like an integration issue between SnapDrive and the vSphere api.

Are you still able to remove the RDM's manually?

Have you posed the question to NetApp?

-KjB

[GOgden]

Yes, within Edit Settings I can remove the RDM manually. Sometimes however it is not obvious where the LUN is connected as it's not visible in the OS so it is removed from the filer, and then the server will crash. Re: the dead FC paths, manually doing a Rescan All on the Datastores will clear these up, but it should happen automatically!

I have also posted this at Netapp https://communities.netapp.com/thread/19192. I wondered if anyone had noticed similar behaviour before I escalate to Support.

Thanks,

Graeme

[Douglas30]

I have the same problem and it is sporadic. Sometimes the LUN will completely clean itself from the host the VM is running on, but the rest of the hosts in the cluster do not recognize the change. Then other times the LUN won't completely remove itself from the host it is running on. Doing a full rescan of the cluster usually sorts this problem out, however I have had to reboot the ESX host to clear this up entirely at one point.

I had a case open with Netapp a while back and they told me it was a VMware issue.

The problem I think is ESX host isn't initiating a cluster rescan after a LUN is removed from the host. Snapdrive removes the LUN from VM, host rescans datastore. At this point the host should tell the cluster to rescan as well, but it isn't.

I'm going to open a case with VMware, see where that gets me. I'll post back.

Unless you alread have an answer back from support?

[GOgden]

Bug 515927 - "Removing a LUN from an ESX host causes multipath software to report that all the paths to the LUN are down." was fixed in SnapDrive 6.4 so we upgraded all servers to that version.

Be very careful though. We subsequently upgraded all our ESXi hosts to 5.0 Update 1 - this completely broke the management service and we were unable to remove/add LUNs or vMotion VMs. NetApp support confirmed that their "fix" in 6.4 was infact a workaround, and that something in ESXi 5.0 Update 1 means this workaround no longer works. They wouldn't confirm any plans to resolve this in 6.4.1 which is currently in beta and suggest we contact VMware to find out what has changed and if it will be resolved in Update 2.

We ended up rolling back to ESXi 5.0 without Update 1 and now everything is fine using that with SnapDrive 6.4

[johari123]

Hi Team

I have question.my customer having some problem,the VM used RDM,but disconnected,on edit setting disable and also remove inventory also disabled,any suggestion how to solve this problem..

Thank you