Windows 2008 R2 SP1 Disk Management problem

[donikatz]

I'm migrating a VBR server from W2k3 R2 to W2k8 R2 and am having a hard time figuring out how to get Disk Management to stop asking about initializing unknown disks (i.e. the passive path FC SAN LUNs).

I've tried the usual: diskpart, automount disable, automount scrub. I've also tried mountvol /N, mountvol /R. Both of those are the same, they just set the NoAutoMount registry setting, which does appear to have been properly set to 1. So it "worked".

Despite this, however, Disk Management still pops up asking to initialize the unknown disks. In W2k3 disabling automount stopped that. Googling seems to indicate disabling automount should still be the way to go for W2k8 R2. Yes, I've tried rebooting in between commands.

I'm stumped. Am I missing something? I'm sure someone else running VBR on W2k8 R2 must know the answer. Thanks!

[Vitaliy S.]

Hello Doni,

I've been googling for your problem and it seems like you're not the only one who suffers from this issue. Looks like this behavior can be observed with FC connections, see this topic:
http://communities.vmware.com/thread/268730

Further research led me to this post, you may want to give it a try:
http://communities.vmware.com/thread/298993

Thanks.

[donikatz]

Thanks Vitaliy, I found some of those posts too, but unfortunately no one that gives a real follow-up. I've tried changing the SAN policy to OfflineAll in Diskpart, but it has no impact: it still wants to initialize the unknown disks (which isn't surprising, since Offline vs Online is really a post-initialization status). I'm surprised this isn't a more-common question, considering how crazy dangerous it is. So now I'm terrified our Veeam admins may one day wipe out an entire SAN with an errant click. You guys are Microsoft partners, right? Can you query your MS engineering contacts about this? Maybe there's a registry hack or something. Thanks

[Daveyd]

Thanks Vitaliy, I found some of those posts too, but unfortunately no one that gives a real follow-up. I've tried changing the SAN policy to OfflineAll in Diskpart, but it has no impact: it still wants to initialize the unknown disks (which isn't surprising, since Offline vs Online is really a post-initialization status). I'm surprised this isn't a more-common question, considering how crazy dangerous it is. So now I'm terrified our Veeam admins may one day wipe out an entire SAN with an errant click. You guys are Microsoft partners, right? Can you query your MS engineering contacts about this? Maybe there's a registry hack or something. Thanks

I am interested in this as well. I have a physical Veeam server running 2008 R2 and have iSCSI drives from my HP SAN via a DataCore SANmelody server presented to it in Read Only mode. Disk Management shows them as Read Only. However, the Veeam server also has a Fiber Channel HBA installed which i plan to attach to our Calriion CX-4 and present VMFS LUNs to it. There is no way in hell I want the Veeam server to even think about attempting to write a signature to those LUNs.

[tsightler]

What type of SAN? Sound's like you perhaps simply need to install the appropriate multipathing support for your SAN vendor.

[donikatz]

We're not using MPIO, single path has been enough thus far for the backup servers vs ridiculous EMC Powerpath MPIO costs. For years we've always been able to just connect to the SAN and disable the double-presented disks of the passive path (you know which they are because they are seen as unknown rather than recognized basic disks). If a path fails over, we just manually change which disks are disabled. Rare occasion, takes a few minutes, saves thousands of dollars. In W2k3 this wasn't a problem because disabling autorun stopped Disk Manager from asking to initialize unknown disks. In W2k8 R2, it always asks to initialize these unknown disks until they're disabled, meaning if a path fails over the operator will be asked to initialize SAN disks. That's the very risky change. There's got to be a registry setting or some other way to stop this...

(Davey, if you're using Powerpath with your Clariion and disable autorun, this won't be a problem. I'm not.)

[tsightler]

As far as I know, you can run Powerpath in "unlicensed" mode completely free. This mode provides "failover only" no true multipath load balancing. That's what we always did when we had Powerpath. It's officially only supported with a single HBA, but we never had any problems even with dual HBA's.

Another option would be to simply zone the system to not present the secondary paths to the host. Of course you would have to do some reconfiguration in the event of a failover, but that's true with your current setup as well.

[donikatz]

Uhm, wow. Well that's embarrassing: been using Clariions since 2005, never knew that. Google says you're right, PP unlimited is still free. Guess I'll download the latest and see what happens. Thanks Tom! (It's still odd that W2k8 R2 behaves this way, but I guess that won't matter.)

[tsightler]

It's still odd that W2k8 R2 behaves this way...

I certainly don't disagree, my suggestion are just some possible options to work around what is obviously brain-dead behavior on the point of Windows 2008R2.

[Daveyd]

(Davey, if you're using Powerpath with your Clariion and disable autorun, this won't be a problem. I'm not.)

Let me know how you make out with the "free" PP as I will be using it on my physical Veeam server with 1 FC HBA

[jctong]

this doc solved it for me

http://download.microsoft.com/download/ ... 8/mpio.doc

[Daveyd]

this doc solved it for me

http://download.microsoft.com/download/ ... 8/mpio.doc

Since it is 100 degrees outside and I am real lazy, can you summarize what you did to fix the issue?

[donikatz]

I'm happy to report back that installing PP in unlicensed free mode solved the problem with ease. No configuration necessary since the host was already registered with the SAN via NaviAgent: just installed PP with defaults, rebooted, and everything looks great, only one path is presented to the OS, no unrecognized disks in device manager. I wish I'd known about this free option years ago. Thanks again, Tom!

[jmartin38134]

Wish to revive this topic with an issue I' seeing. Have Clariion CX3 / FC connectivity / Veeam 5.0.2.230(64bit) running on W2K8 R2 Std with PowerPath v5.5 64bit installed. Everything looks good with one path presented in Data Mgmt but I have 33 LUNs presented 32 show Basic/Online/Healthy(Primary Partitioni)) and one shows Basic/Online/Unallocated. When I attempt a Direct SAN Backup of a VM on the LUN showing Unallocated it fails to do direct backkup and revertst to NBD backup. Does anybody know what causes a disk to show Unallocated? What W2k8R2 using to determine Healthy?

[foggy]

James, could it be a kind of PowerPath limitation to support up to 32 paths to a device? You can probably "replace" one LUN with another and check if the healthy LUN becomes Unallocated and vice versa.

[jmartin38134]

LUN showing Unallocated. I am on the way to resolving this. It is a VWware problem. If I go to a ESX host Configuration Tab / Storage and right click the datastore that shows unallocated then click properties, there are no Primary Partitions. All other datastores show a Primary Partition simular to
Primary Partition Capacity
1. VMFS 350.00 Gb

Will resolve with VMware.

Thanks.