SQL Server 2008 SP1 installation has failed on Windows 2008 single node cluster

  • Hello,

    We are installing SQL Server 2008 x64 on Windows 2008 x64 Single node cluster

    Steps performed:

    1.Created the Cluster on Single node on Windows 2008 x64 with SP2

    2.Assigned the below drives

    C: OS –Local drive

    D: To place the SQL Server binaries – Local drive

    F: Primary data files – Shared drive

    G: Secondary data files - Shared drive

    H:Log files - Shared drive

    I: TempDB - Shared drive

    Z: Backup Drive - Shared drive

    Q: Quorum - Shared drive

    Y: MSDTC - Shared drive

    3.Created and configured MSDTC

    4.Have the QUORUM drive but did not see any option to assign the QUORUM. I believe, QUORUM will be assigned when add the second node (correct me if we miss anything here)

    5.Added the Admin account to the Administrator Group and Started the SQL Server 2008 Installation

    6.The SQL Server 2008 installation has failed with below error.

    The following error has occurred:

    The cluster resource 'SQL Server (INS2)' could not be brought online. Error: The resource failed to come online due to the failure of one or more provider resources. (Exception from HRESULT: 0x80071736)

    7.I noticed the below error in system event viewer errors:

    a)

    Log Name: System

    Source: Microsoft-Windows-FailoverClustering

    Date: 8/20/2010 10:10:12 AM

    Event ID: 1194

    Task Category: Network Name Resource

    Level: Error

    Keywords:

    User: SYSTEM

    Computer: SQL2.abc.com

    Description:

    Cluster network name resource 'SQL Network Name (SQLVS)' failed to create its associated computer object in domain 'abc.com' for the following reason: Unable to create computer account.

    The text for the associated error code is: Access is denied.

    Please work with your domain administrator to ensure that:

    - The cluster identity 'SQLCLUSTER$' can create computer objects. By default all computer objects are created in the 'Computers' container; consult the domain administrator if this location has been changed.

    - The quota for computer objects has not been reached.

    - If there is an existing computer object, verify the Cluster Identity 'SQLCLUSTER$' has 'Full Control' permission to that computer object using the Active Directory Users and Computers tool.

    b)

    Log Name: System

    Source: Microsoft-Windows-FailoverClustering

    Date: 8/20/2010 10:10:12 AM

    Event ID: 1069

    Task Category: Resource Control Manager

    Level: Error

    Keywords:

    User: SYSTEM

    Computer: SQL2.abc.com

    Description:

    Cluster resource 'SQL Network Name (SQLVS)' in clustered service or application 'SQL Group' failed.

    8.Uninstalled SQL Server 2008

    9.Then Windows Admin created the Virtual Server name SQLVS in Active Directory and granted the permission to create objects in DNS

    10.Installed the SQL Server 2008 again and it was successful

    Questions:

    1.Why the SQL Server 2008 setup unable to create the Virtual Server name SQLVS in Active Directory? Is this by design?

    2.Do we need to create the Virtual Server name in Active Directory and granted the permission to create objects in DNS prior to install SQL Server 2008?

    11.Started the installation of SQL Server 2008 SP1 but it’s failed

    12.The resources went offline and I started them manually and they are online

    13.Below is the error from Summary log file:

    Overall summary:

    Final result: The patch installer has failed to update the following instance: INS2. To determine the reason for failure, review the log files.

    Exit code (Decimal): -2068709375

    Exit facility code: 1202

    Exit error code: 1

    Exit message: The patch installer has failed to update the following instance: INS2. To determine the reason for failure, review the log files.

    Start time: 2010-08-21 14:06:42

    End time: 2010-08-21 14:09:55

    Requested action: Patch

    Instance INS2 overall summary:

    Final result: The patch installer has failed to update the shared features. To determine the reason for failure, review the log files.

    Exit code (Decimal): -2068709375

    Exit facility code: 1202

    Exit error code: 1

    Exit message: The directory 'g:\baee41947d8ca4d0626e10\' doesn't exist.

    Start time: 2010-08-21 14:09:07

    End time: 2010-08-21 14:09:53

    Requested action:

    Questions:

    1.Why the SP1 is failing to update Shared features?

    2.Why the SP1 setup reporting that the directory G: does not exist?

    3.What will be resolution to fix this and install SP1 successfully?

    I tried SP1 installation several times but it’s failing with the same error. But each time it’s report different Drive as it’s does not existing. I mean not G: drive every time

    1st time, it’s reported G: drive does not exist

    2nd time, it’s reported H: drive does not exist

    Note: While installing SP1, I noticed two warnings as below and I have ignored them. Please advice if we need to fix them before installing SP1

    a) Cluster group owned by local computer (here, as we have single cluster node, we can ignore this warning right?)

    b) Cluster resource DLL update restart check

    Please advice. I appreciate your inputs

    Thanks

  • Hi,

    I have followed the below link where same issue has discussed.

    http://sqlserverpedia.com/blog/sql-server-bloggers/failure-when-upgrading-cluster-to-sql-2008-sp1-cu7/

    As Shaun Stuart, suggested, I have stopped the SQL Server service and the Agent before installing SQL Server 2008 SP1. But still I'm getting the same error:

    Instance INS2 overall summary:

    Final result: The patch installer has failed to update the shared features. To determine the reason for failure, review the log files.

    Exit code (Decimal): -2068709375

    Exit facility code: 1202

    Exit error code: 1

    Exit message: The directory 'f:\86014e80ae985af1632d7978e8b895\' doesn't exist.

    Start time: 2010-08-21 17:46:26

    End time: 2010-08-21 17:47:11

    Requested action: Patch

    Now the SQL Server instance is NOT coming online

    Looks like the instance is messed up:hehe:

    Is this a bug? Do we have any known workaround?

    Please advice

    Thanks

  • pshaship (8/21/2010)


    I believe, QUORUM will be assigned when add the second node (correct me if we miss anything here)

    no, no , no , no , no. The quorum is present now, even a single node cluster has to have the quorum. It is quite possibly using file share

    In the Windows 2008 Failover cluster manager click on the cluster name in the left pane. The centre pane will display cluster configuration information the quorum type will be listed there. Please post back the type it indicates you are using

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • In the Windows 2008 Failover cluster manager click on the cluster name in the left pane. The centre pane will display cluster configuration information the quorum type will be listed there. Please post back the type it indicates you are using

    Thank you,

    Here is the configuration from Failover cluster manager:

    Quorum Configuration: Node Majority - Warning:failure of node will cause the cluster to fail. Check the status of nodes

    Here, we are dealing with Single node

    Thanks for your help

  • As i suspected you are using file share quorum.

    To recap, these are the quorum models in Windows 2008 clustering

    Node Majority—There's no file-share witness or disk witness and votes are assigned only to nodes. More than 50 percent of the nodes must be available to make quorum. Node majority works best if you have an odd number of nodes.

    Node and Disk Majority—Nodes and a shared disk get votes. This configuration allows a loss of half the nodes, providing the disk witness is available, or over half the nodes are available without the disk witness being available. Recommended if your cluster has an even number of nodes.

    Node and File Share Majority—The same as Node and Disk Majority, except the shared disk is replaced with a file-share witness. Recommended if you have multi-site cluster, an even-node cluster, a cluster with no shared storage, or a two-node cluster.

    No Majority: Disk Only—The disk witness must be available to have quorum, but the number of available nodes doesn't matter. If you have a four-node cluster and only one node is available, but the disk witness is available, you have quorum. If the disk witness isn't available, then even if all four nodes are available you can't have quorum. No longer recommended to have the No-Majority: Disk-Only mode.

    Personally, I opt for the Node and File Share Majority eliminating the need for an extra disk resource.

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • As i suspected you are using file share quorum.

    NO. As you see my Quorum configuration says Node Majority.

    Quorum Configuration: Node Majority - Warning:failure of node will cause the cluster to fail. Check the status of nodes

    To recap, these are the quorum models in Windows 2008 clustering

    Node Majority—There's no file-share witness or disk witness and votes are assigned only to nodes. More than 50 percent of the nodes must be available to make quorum. Node majority works best if you have an odd number of node

    .

    The SQL Server 2008 installation is a success but NOT the SP1. So how does this Quorum cause the SP1 to fail but not SQL Server 2008.

    What is the logic? I did not get it

    Thank you

  • pshaship (8/22/2010)


    As i suspected you are using file share quorum.

    NO. As you see my Quorum configuration says Node Majority.

    :blush: sorry my bad. yes. Was meant to say Node majority

    pshaship (8/22/2010)


    The SQL Server 2008 installation is a success but NOT the SP1. So how does this Quorum cause the SP1 to fail but not SQL Server 2008.

    What is the logic? I did not get it

    Thank you

    Quorum will not affect the SP1 install.

    pshaship (8/22/2010)


    a) Cluster group owned by local computer (here, as we have single cluster node, we can ignore this warning right?)

    the patching is usually carried out on passive nodes first, but you only have one node anyway so it should force a reboot.

    Is there anything else in the setup log?

    did you shutdown the clustered service and take all drives offline? This should force setup to use another disk

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • sorry my bad. yes. Was meant to say Node majority

    So having Node majority quorum makes sense for Single node cluster?

    There are NO errors except the below for SP1 failure:

    I tried placing the software on Local disk and then install SP1 & then Stopping the SQL Server services and install SP1. But still SP1 failing with same error. Any help would be appreciated!

    Instance INS2 overall summary:

    Final result: The patch installer has failed to update the shared features. To determine the reason for failure, review the log files.

    Exit code (Decimal): -2068709375

    Exit facility code: 1202

    Exit error code: 1

    Exit message: The directory 'f:\86014e80ae985af1632d7978e8b895\' doesn't exist.

    Start time: 2010-08-21 17:46:26

    End time: 2010-08-21 17:47:11

    Requested action: Patch

    Thanks

  • What I noticed is the SP1 installer is creating some temporary folder and installation files in it. In the middle of the installation all drives are going offline & at that time SP1 installation fails. So for some reason the setup unable to bring the disk resources online and there fore loosing access to the temporary folder & install files and it;s fails there.

    Please see the attachment for the Temp folder and it's install files created by SP1 during the installation:

    Do we need to remove dependencies from SQL Server (instance). i.e the disk resources before installing SQL Server 2008 SP1?

    What are the steps to perform prior to SP1 installation in single node cluster?

    Please advice

  • Taking the SQL server service resource offline will not offline the disks. The installer must be offlining the whole group.

    Take the whole group offline before running the installer, not just the SQL server service.

    Relaunch installer and try again!

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • just a thought.

    Have you tried manually extracting the SP1 to a temp folder on C drive and launching it from there??

    execute the following

    en_sql_server_2008_sp1_x64.exe /X:"C:\SP1"

    Once the files are extracted launch the service pack from

    C:\SP1

    Let me know how you get on

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • Hi Perry,

    You are awesome. Extracting the SP1 to temp folder & launching from there is worked:-)

    But still I would like to know the reason why SP1 is unable to extract files itself in Single Cluster node (Windows 2008)?

    Thanks so much!

  • pshaship (8/24/2010)


    Hi Perry,

    You are awesome.

    ha ha 🙂 Dude you are too kind!

    pshaship (8/24/2010)


    But still I would like to know the reason why SP1 is unable to extract files itself in Single Cluster node (Windows 2008)?

    Thanks so much!

    We know the reason, the whole group was going offline including the disks! I've never had the problem myself even with a single node but then i always slipstream SP1 whilst installing the clustered instance.

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

  • We know the reason, the whole group was going offline including the disks! I've never had the problem myself even with a single node but then i always slipstream SP1 whilst installing the clustered instance.

    If Microsoft knows that the whole group is going to offline in a Single cluster node, then why Microsoft did not provide the documentation that we MUST need to Slipstream SP1 in Single node cluster?

    If you are NOT going to use Slipstream, then why Microsoft did not provide the documentation that we need to manually extract the SP1 files & then install ?

    Just to make a reasonable argument?

    Thanks

  • Probably because a cluster with a single node defeats the object of clustering, there is no failover!

    -----------------------------------------------------------------------------------------------------------

    "Ya can't make an omelette without breaking just a few eggs" 😉

Viewing 15 posts - 1 through 15 (of 18 total)

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