SQL2K/Cluster installation

  • I'm installing SQL2K EM on a IBM cluster. I'm getting "Setup failed to perform the required operation on the cluster nodes". I do look around Microsoft KB and found some documents but none of them apply to my error. In the setup log, I saw some error message

    ShowDlgDiskGroups: GetLastError returned: 1008

    An attempt was made to reference a token that does not exist.

    I knew it failed when try to do the remote installation on the secondary node. But, I have no clue where to check now. The ID I use to install does have local admin permission on both nodes. I also try installation from either node, it all had same problem. I test fail-over cluster on O/S level and also seems working fine.

  • Do your disk groups have spaces in them? If so try removing them and try again. Also check that you have WINS entries on both machines.

    Finally, if it still doesn't work, check the server local security policies and then post the applicable part of the setup log so that we can see the errors.



    Shamless self promotion - read my blog http://sirsql.net

  • It is brand new server has C: 8GB, D: 144 GB and E: 144 GB. Only O/S is on. I do not believe there is any space, special characters on any naming convention.

    Sorry for my lack of knowledge. How do I verify there is WINS entries ?

    Since I knew the error is from remote instllation on remote node, below is the error from installing node (PRODSQL02):

    Installing remote service (PRODSQL01)...

    CProcess::RunUntilComplete [0x94] exit code: 2

    Remote process exit code was '2' (PRODSQL02).

    CThreadPool::RunUntilCompleteHlpr WaitForMultipleObjects returned: 0

    CThreadPool::RunUntilCompleteHlpr signaled thread [0x94]

    Thread [0x94] exit code: [0x2]

    Running '\\PRODSQL02\Z$\x86\setup\setupsql.exe k=ClSec k=Rm k=Cl -SMS -s -f "\\PRODSQL02\Z$\x86\setup\setup.ins" -f1 \\PRODSQL01\ADMIN$\PRODSQL01_MSSQLSERVER.iss -f2 "\\PRODSQL01\admin$\setup.log" -e "stpsilnt._ex" -x "C:\"' (PRODSQL01) ...

    CRemoteProcess::RunUntilComplete [0x98] exit code: 2

    Remote process exit code was '2' (PRODSQL01).

    CThreadPool::RunUntilCompleteHlpr WaitForMultipleObjects returned: 2

    CThreadPool::RunUntilCompleteHlpr signaled thread [0x98]

    Thread [0x98] exit code: [0x2]

    CThreadPool::RunUntilComplete returned 2

    CThreadPool::RunUntilComplete execution level=1, need execution: 0

    One or more errors occurred while running the remote/unattended setups.

    Disconnecting from remote machine (PRODSQL02)...

    Service removed successfully.

    Remote files could not be removed.

    #### SQL Server Remote Setup - Stop Time 04/12/04 20:35:03 ####

    Disconnecting from remote machine (PRODSQL01)...

    Service removed successfully.

    Remote files removed successfully.

    #### SQL Server Remote Setup - Stop Time 04/12/04 20:35:15 ####

    .....

    20:35:15 End Action : Run setup on cluster nodes

    20:35:15 One or more errors occurred while running remote setup on the clustered machines.

    20:35:29 Setup failed to perform required operations on the cluster nodes.

    20:35:29 Begin Action : ClusterErrorCleanup

    20:35:29 End Action : ClusterErrorCleanup

    20:35:29 End Action : ClusterSetupHelper

    Error from remote node (PRODSQL01)

    ...

    20:35:00 [DlgVirtServer]

    20:35:00 Result = 1

    20:35:00 VirtServerName = PRDSQL

    20:35:00 IpAddresses = 10.100.100.47,255.255.255.0,public

    20:35:03 ShowDlgVirtualServer returned: 0

    20:35:03 Ipaddress = 10.100.100.47,255.255.255.0,public

    20:35:03 szNewList = 10.100.100.47,255.255.255.0,public

    20:35:03 szModifiedList =

    20:35:03 szRemovedList =

    20:35:03 End Action DialogShowSdVirtualServer

    20:35:03 begin ShowDialogsUpdateMask

    20:35:03 nFullMask = 0x1bf7c037, nCurrent = 0x4000, nDirection = 1

    20:35:03 Updated Dialog Mask: 0x1bf7c037, Disable Back = 0x1

    20:35:03 Dialog 0x4000 returned: 1

    20:35:03 End Action ShowDialogsHlpr

    20:35:03 ShowDialogsGetDialog returned: nCurrent=0x8000,index=15

    20:35:03 Begin Action ShowDialogsHlpr: 0x8000

    20:35:03 Begin Action : DialogShowSdDiskGroups

    20:35:03 [DlgDiskGroups]

    20:35:03 Result = 1

    20:35:03 DiskGroupsName = Disk

    20:35:03 ShowDlgDiskGroups returned: -1

    20:35:03 ShowDlgDiskGroups: GetLastError returned: 1008

    20:35:03 An attempt was made to reference a token that does not exist.

    20:35:03 End Action : DialogShowSdDiskGroups

    20:35:03 End Action ShowDialogs

    20:35:03 Action CleanUpInstall:

    20:35:03 StatsGenerate returned: 2

    20:35:03 StatsGenerate (0x0,0x1,0xf00000,0x400,1033,0,0x0,0x1000000a,0,0,0

    20:35:03 StatsGenerate -1,sql_svc)

    20:35:03 Installation Failed.

  • Try renaming the Disk Group, also are you using an account that has administration rights on the cluster itself?

    You can check WINS entires under the TCP/IP settings on the public network connection of your cluster.



    Shamless self promotion - read my blog http://sirsql.net

  • Is the SQL CD in the server (primary node) or are you mapping to  a network? Mapping to a network CD caused those errors for me in the past.

  • Nope, it is local. I also try copy the whole CD to the local C: to do the install. Same problem.

  • 1. if you haven't done that make sure that unnecessary services are stopped

    2. check the cluster logs for an error message regarding gathercl.dll. If you have it you will have to manually register it on %windir% \cluster on both machines

    3. try to install the SQL Server cluster from the other node (not the one that you have already used)

  • I am having this very same issue. Did anyone ever find a solution to this problem? What were the steps?

     

    Thanks

  • Are you using IBM SCSCI H/W ?

  • No, HP 4m RAID Disk Set.

  • SQL Server Clustering requires the disks that are used by SQL to be available to all nodes.  It looks like drives D: and E: might be.  These disks will need to be managed by the cluster so that when SQL Server fails over to the other node, the cluster manager will give the second node access to them and SQL can be started.

    Normally, only one node is allowed to see the disks because SQL Server clusters share no resources.  Hence, if you run windows explorer on each of the nodes, only one of them will be able to see the disks.

    Hope this helps

    Stephen

     

  • In general, the message indicate somewhere the cluster is not setup correctly or something wrong on the H/W (most possible is RAID controller). In our situation, after replace the RAID controller and apply a patch to the RAID, the problem went sway.

Viewing 12 posts - 1 through 11 (of 11 total)

You must be logged in to reply to this topic. Login to reply