locked
Windows 2008 cluster disk reserved RRS feed

  • Question

  • I have 2 Windows Server 2008 servers and I'm using a Dell Equallogic SAN.  these 2 servers are setup as a cluster.  I originally configured them a few  months ago and then I didn't touch them for about 2 months.  Now I went back to them and all of the disks are showing up as reserved on both servers.  There's no options in disk management, everything is greyed out. 

    I have tried to take the SAN offline, bring it back online and then bring the servers up 1 at a time.  that didn't help. 

    It's the same for all of the SAN disks.  I contacted Dell Equallogic and they do not see a reservation on the SAN.  They asked me to contact Microsoft. 

    I verified that the Cluster service is running. 

    I also tried using "cluster node w2k8 /clear:3" where 3 is the number of the disk.

    I also tried running the Validation tests again. 

    Thursday, March 26, 2009 4:10 PM

Answers

  • Found answer in another forum:

    Sounds like stale PR....easiest is to use cluster.exe CLI to try clearing the PR. Here is a sample command

    cluster node w2k8-cl1 /clear:3

    3 = the number of the disk as seen in the disk Management interface.

    You can also re-run all the validation storage tests ensuring you clear the check box in the wizard allowing the validation to run against all storage. Part of the validation process for storage is to clear all PRs on the drives.
    ---------------------------------------------------------------

    After clearing the reservations, I went to Cluster Management and 2 of the disks showed online, the 3rd one showed offline and I was able to select the disk and put it online.

    Then I ran the validation tests again and they all passed.

    Apparantely it's normal for Windows 2008 disks to show up as reserved, according to Microsoft.

    You can stop the cluster service in Failover Management on one of the nodes and then the other server should take over the drives. If you start the cluster service, that server will not automatically take over until you stop the cluster service on the other node.
    • Marked as answer by JohnToner Thursday, July 29, 2010 7:43 PM
    Friday, April 3, 2009 2:52 PM

All replies

  •  What is the status of the disk resources in Cluster Administrator?
    Visit my blog about multi-site clustering - http://msmvps.com/blogs/jtoner
    Thursday, March 26, 2009 4:34 PM
  •  What were the results of the two things you tired?  Especially the validation tests.
    Chuck Timon Senior, Support Escalation Engineer (SEE) Microsoft Corporation
    Thursday, March 26, 2009 6:05 PM
  • Ok, after running the validation tests, here's what FAILED:

    1.  NETWORK
    -Validate IP configuration
    Found duplicate IP address fe80::100:7f:fffe%12 on node mail3.test.com adapter Local Area Connection* 11 and node mail4.test.com adapter Local Area Connection* 11.


    Local Area Connection 11 on MAIL4
    Adapter Name Local Area Connection* 11
    Adapter Description Teredo Tunneling Pseudo-Interface
    Physical Address 02-00-54-55-4E-01-00-00
    Status Not Operational
    IP Address fe80::100:7f:fffe%12
    Prefix Length 64 

    Local Area Connection 11 on MAIL3
    Adapter Name Local Area Connection* 11
    Adapter Description Teredo Tunneling Pseudo-Interface
    Physical Address 02-00-54-55-4E-01-00-00
    Status Not Operational
    IP Address fe80::100:7f:fffe%12
    Prefix Length 64


    I have no idea what the Teredo Pseudo interface is. 

    2.  STORAGE

    Validate File System
    Failed to online cluster disk 0 from node mail3.test.com, failure reason: The requested resource is in use.
    An error occurred while executing the test. The string is null or empty. Parameter name: message

    Validate Simultaneous failure
    Failed to arbitrate for cluster disk 0 from node mail3.test.com, failure reason: The requested resource is in use

    ---------------------------------------------------------------------------------
    After using the CLUSTER command, on Mail3 I'm able to access 2 of the disks by going to My Computer.  However, 2 of the disks show up as failed in Disk Management and all 3 disks show up reserved.

    If I go to Cluster Management, it shows as only 1 disk being OFFLINE on Mail3.  The other disks show up as being Online. 


    Under Cluster Events, i see:

    Event ID 1069
    Source: Failover clustering
    Cluster resource 'Cluster Disk 3' in clustered service or application 'Cluster Group' failed.

    Event id 1037
    Source: Failover clustering
    The file system for one or more partitions on the disk for resource 'Cluster Disk 3' may be corrupt. Run the Validate a Configuration wizard to check your storage configuration. Optionally, you may want to run Chkdsk to verify the integrity of all volumes on this disk.

    Event id 1038
    Source: failover clustering
    Ownership of cluster disk 'Cluster Disk 3' has been unexpectedly lost by this node. Run the Validate a Configuration wizard to check your storage configuration.





    Friday, March 27, 2009 11:56 PM
  • For Teredo issue, see
    http://blogs.msdn.com/clustering/archive/2008/07/26/8773796.aspx

    For your disks, delete the failed/offline disks from cluster adminstrator. This will cause them to stop showing up "reserved" but of course does not necessarily resolve your problem. Once you've taken the disks out of cluster, run thru the validation again for these disks and see if it passes. If not, you'll likely need to engage your storage vendor.


    Visit my blog about multi-site clustering - http://msmvps.com/blogs/jtoner
    Monday, March 30, 2009 9:08 PM
  • Found answer in another forum:

    Sounds like stale PR....easiest is to use cluster.exe CLI to try clearing the PR. Here is a sample command

    cluster node w2k8-cl1 /clear:3

    3 = the number of the disk as seen in the disk Management interface.

    You can also re-run all the validation storage tests ensuring you clear the check box in the wizard allowing the validation to run against all storage. Part of the validation process for storage is to clear all PRs on the drives.
    ---------------------------------------------------------------

    After clearing the reservations, I went to Cluster Management and 2 of the disks showed online, the 3rd one showed offline and I was able to select the disk and put it online.

    Then I ran the validation tests again and they all passed.

    Apparantely it's normal for Windows 2008 disks to show up as reserved, according to Microsoft.

    You can stop the cluster service in Failover Management on one of the nodes and then the other server should take over the drives. If you start the cluster service, that server will not automatically take over until you stop the cluster service on the other node.
    • Marked as answer by JohnToner Thursday, July 29, 2010 7:43 PM
    Friday, April 3, 2009 2:52 PM
  • try to disable the firewall,
    Friday, August 5, 2011 2:32 PM
  • Above Steps work perfect for Me.

     

    Thanks for the solution :)

    Monday, December 26, 2011 10:17 AM
  • I seem to be getting this error, only I get it every single time I reboot one of the nodes. Each time, it seems like there's some kind of PR check or something - 2 node cluster, just a test shared disk in the middle. Nodes ride on top of ESXi 5.1, shared disk is RDM. Not even formatted yet, just initialized. Sun/Oracle Open Storage 7320 clustered SAN using FC.

    Any time I reboot a node, the disk immediately goes offline. If I shut down, it goes offline. If I stop cluster service, ownership slides over just fine, just as expected. But shutting down kills the disk. If I leave it off, and re-online the disk it'll stay online indefinitely. If I then start the other node, immediately goes offline - during POST, seems like. I can online it manually, and again - runs fine. Cluster node joins up, seems to be healthy. Everything passes all validation checks.

    I'm baffled.

    Saturday, October 27, 2012 12:16 AM
  • For Windows 2012, if the cluster.exe commands don't fix it, this method is an option:

    Regedit.exe

    HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Services\ClusDisk\Parameters

    AttachedDisks

    This string has the Persistent Reservation disk signatures identified.  Export this key for backup and then delete AttachedDisks from the registry.  Reboot.

    Now in Disk Management the reserved flag should be gone.  Repeat on other nodes as needed.  Make sure multiple nodes don’t try to access the disk while not clustered!  Leave them offline until sure there is no simultaneous access.

    Wednesday, January 16, 2013 9:55 PM
  • It looks like scsi 3 reservation problem.

    If you do the same thing with windows 2003 you don't see this problem

    Thursday, June 13, 2013 12:37 PM
  • If your disks shows as Reserved then the Disk are been Controlled by Cluster. You need to Go to the Cluster Console and navigate to Storage Tab in The Cluster Storage and from there you can the disks are been added on cluster. IF they are in Available Storage Section then you can just right click and remove them from Cluster Control and you are done they can used after that by LDM.

    On Cluster Disk Driver and Cluster Driver Structure then have look on : http://aikitsupport.com/how-to-disable-cluster-disk-driver/

    Friday, June 14, 2013 12:58 PM
  • We spoke oracle  engineering team . They said it is the bug and they are changing it. In the next release you won't have this problem

    VMware told us to set the perennially reserved on all the RDM LUNS.

    http://kb.vmware.com/kb/1016106

    Friday, June 28, 2013 7:13 PM