How to create SQL Server 2005 Clustering with Virtual Server 2005 (Part 4)

How to create SQL Server 2005 Clustering with Virtual Server 2005 (Part 4)

Rate This
  • Comments 20

In previous parts(Part 1, Part 2 and Part 3), we have completed windows clustering. Now, before installing SQL Cluster we have to configure MSDTC. We also need to make some shared drive which we would use to keep SQL Server databases and MSDTC. We are going to use Disk M for MSDTC and Disk X for SQL Server. We have already done the steps for Disk Q in part 2 (Create Quorum Disk section). I will reproduce the steps for Disk M and X here.

Configure Shared Disks

1. Turn off MyNode1 and MyNode2

2.  Choose Virtual Disks > Create > Fixed Size Virtual Hard Disks

Create Virtual Disk(M Drive)

3. Provide Details for M Drive (250 MB) and hit "Create" button

Create Virtual Disk(M Drive) Details

4. Same way we would create X drive (2 GB) and Hit "Create" button.

Create Virtual Disk(X Drive) Details

Create Virtual Disk(Y Drive) Details

5. Below is what you would see in Windows Explorer

All Disks

6. Lets attach the disks to MyNode1 and MyNode2 (They are off at this moment) I am showing for MyNode1. Please follow same for MyNode2 also.

MyNode1_Edit Configuration

 MyNode1 Edit hard disk

Click on Hard Disks

MyNode1 HardDisks Properties

Hit OK and it would attach the disks to MyNode1. Follow the same for MyNode2

7. Turn on ONLY MyNode1.

8. Login to MyNode1 and Start > Run > DiskMgmt.msc. Cancel on the screen if you get any.

9. You should see two new Disks which are Not Initialized. Right Click and choose Initialize Disk.  Then you create partition by Right Click New Partition. Below is final screen shot which you should see.

MyNode1 Disk management

10. Now Turn Off MyNode1 and Turn On MyNode2

11. Go to DiskMgmt.msc on MyNode2 and assign drive letters (M for 250 MB drive and X for 2 GB Drive)

12. Open Cluster Administrator (Start > Run > CluAdmin)

13. Create New Group called MSDTC (File > New > Group) and provide details

New Group MSDTC

14. Choose Both Nodes as preferred owner. You might get an error at last screen but that's OK because MyNode1 is offline.

15. Create a new "Disk" resource under MSDTC Group (File > New > Resource) Choose details as below screen shot.

New Resource M Drive

17. Choose M Drive in "Disk Parameters" screen. and finish.

Disk Parameters

18.  Now you should see a Disk resource under MSDTC Group. Right click and choose "Bring Online"

19. Follow 13 to 18 again to Create new groups called "SQL Group" and add "Disk X"

SQL Group Online

20. Up to here we have completed shared disk configuration for MyNode1 and MyNode2. Let's start MyNode1 and test fail over for new groups.

21. Before we install SQL Server, we have to configure MSDTC

Configure MSDTC as Resource

  1. On MyNode1 and MyNode2 follow below steps
    • Open Control Panel on the cluster node
    • Double-click Add/Remove Programs
    • On the left side of the screen, double click Add/Remove Windows Components
    • Under Components, highlight Application Server and click Details
    • Select Enable network DTC access, and then click OK
    • Click Next
    • Click Finish
  2. Add "MSDTC IP Address resource" in MSDTC Group
    • Right click MSDTC Group and select New > Resource
    • Type a some name such as MSDTC IP Address
    • In Resource Type, choose IP Address
    • Click Next
    • Click Next on Possible Owners screen
    • Click Next on Dependencies Screen
    • Give IP Address as 192.1.1.5
    • Click Finish
    • Right click the MSDTC IP Address resource and select Bring Online.
  3. Add "MSDTC Network Name" resource in MSDTC Group
    • Right click MSDTC Group and select New > Resource
    • Type a some name such as MSDTC NN
    • In Resource Type, choose Network Name
    • Click Next
    • Click Next on Possible Owners screen
    • On Dependencies Screen, move MSDTC IP Address and click on Add to bring on Right window. Now hit Next.
    • Give Name as MSDTC
    • Click Finish
    • Right click the MSDTC NN resource and select Bring Online.
  4. Add "MSDTC" resource in MSDTC Group
    • Right click MSDTC Group and select New > Resource
    • Type a some name such as MSDTC
    • In Resource Type, choose Distributed Transaction Coordinator
    • Click Next
    • Click Next on Possible Owners screen
    • On Dependencies Screen, choose "MSDTC NN" and "M Drive" and click on Add to bring on Right window.

                        MSDTC Dependencies

    • Click Finish
    • Right click the MSDTC resource and select Bring Online.

MSDTC Group Online

Above is the screen shot of "MSDTC group". Make sure you have matching "Resource Type" as shown in the screen.

Test fail over for MSDTC group. This finishes all Prerequisites to install SQL Server on Cluster. Get ready to install SQL Server in last part of the series.

You can follow http://support.microsoft.com/kb/899191 to make changes to MSDTC Services.

Lets move on to last part of the series

Leave a Comment
  • Please add 7 and 3 and type the answer here:
  • Post
  • After step 11 you need to start up Node 1. I am not sure if this is coorect but worked for me. When i first did it node 1 was shut down. after creating groups and resources, i bought up node 1 and then to simulate failover i disconnect both the networks from node 2. I cluadmin on node 1 there was an error to bring M Drive online, but S drive though was online. So i again connected the newtowrk on node 2 and then restarted the cluster service on node 2. Delete the resource M Drive and then recreated when both node were online. Now i again disconnected network on node 1 and then M drive was up on node but now the S Drive failed (i used S instead of X), i again repetead the step to delelte the resource and then recreate it. After doing so i could simulate failover and all the resources were online.

    Please let me know if you think the process is not correct.

  • Sorry, but I could not understand your explaination...

  • As u said in the first part about to download virtual hard disk, the link having the following three downloads which one is need to be downloaded.

    WIN2K3R2EESP2.part1.exe

    700.0 MB

    Download

    WIN2K3R2EESP2.part2.rar

    700.0 MB

    Download

    WIN2K3R2EESP2.part3.rar

    495.1 MB

  • Hi Thanga,

    Thanks for the comments.

    You have to download all the files in same folder. Once you run part1.exe, it automatically use rest two and extract vhd. hope this answers your question. Feel free to add commments.

    I have modified Part 1 to make it clear.

    - Balmukund

  • I see a lot of mention in this part and part 5 of virutal names. MSDTC requires it's own name and IP. So does SQL server.

    The cluster has a name and for that to work you need a computer account in Active directory. usually the wizard creates this.

    My question is what does adding a virtual name create? Do you need rights in active directory for that name to register. What if you don't have those rights? Does your admin need to create aliases? Does the cluster service account need any rights to them.

    I've been searching high and dry for any information on not being an admin and virutal server names with no luck.

    Thanks

  • Hi Eric,

    When you create any Virtual Network Name in Cluster, it does not create the object in AD. There is a check box  "Enable Kerberos Authentication" at Right Click > Properies > Parameters. If that box is checked, it would registry Virtual Network Name in Active Directory.

    By default an account can create 10 objects in AD.

    Refer - http://support.microsoft.com/kb/307532/en-us

    Please let me know if your question is unanswered.

    - Balmukund

  • Greate article,

    After step 11, I cannot open SQLCluster from Node1 or Node2. If i try to ping sqlcluster.mydomain.com the ping fails. I noticed Cluster service is started and running on both nodes. If i try to create a new cluster and provide the name of sqlcluster, I cannot proceed as sqlcluster can be seen. Any help would be appreciated.

    Regards,

    Paul

  • More information included...

    The error message that i receive is "The Cluster Service on node 'sqlcluster' cannot be started. The network path was not found. Error ID: -2147024843 (80070035).

    Thanks again,

    Paul

  • Paul,

    Sorry to hear that you could not proceed smoothly. Can you please let me know where exactly you faced failure? You mentioned step 11, it would help if you can tell which part of series (Part 3)?

    I would go by looking at event log to check the failure. Is your cluster service running? (your two post contradict each other) If service is not running then check event log to see the reason of failure.

    Please do post back and let us know the problem. I would be happy to help.

  • It was network related issue. I am past that stage now. Thanks for the quick reply.

    Regards,

    Paul

  • Hi,

    After step 21, I tried to test failover for MSDTC group. The other three resources are online but MSDTC resource is offline. How can I correct for this?

    If this MSDTC group not online, will SQL 2005 installation fail?

    Thanks

  • Hi Linda,

    I feel that MSDTC is not configured properly on both the nodes causing it to fail.

    (Configure MSDTC as Resource - Step 1 to be done on both the nodes.)

    Can you please have a look at eventlog on node2 and check the error?

    Run below command on node2

    msdtc -uninstall

    msdtc -install

    and try..

    Hope this helps!

    Balmukund

  • Hi Balmukund,

    I am stuck on step 12 in part 4 series of configuring shared disks -

    12. Open Cluster Administrator (Start > Run > CluAdmin)

    13. Create New Group called MSDTC (File > New > Group) and provide details

    I am not able to open a connection to SQLCluster. I says network path not found.

    Please help.

  • Hi Balmukund,

    I am past that problem. I discovered that cluster services were not turned on. After I turned on the cluster service it started working fine.

    I have another problem ...

    While configuring MSDTC resource -> adding a distributed transaction coordinator it says that it cannot create a duplicate resource of this type.

    Please help.

  • Hi Balmukund,

    I was able to successfully implement everything till step: 20 of part 4.Before going to configure MSDTC i.e step 21.I was testing some of the fail over scenario's between the groups and nodes will work..It looks like moving the Groups/Resources works between the nodes.But when i manually turn off Node1, Now i am unable to access cluster Administrator and i see Drives Q ,M & X are available on Node 2. Then later i bring back the Node1 and hoping for the ClusterService  to be up and running and expecting the failback from Node2 to Node1..Which is not happening..What i am missing? Any body pl's help me....Here is the error from Node1 event log:

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

    Event Type: Warning

    Event Source: MSDTC Client

    Event Category: Cluster

    Event ID: 4148

    Date: 12/21/2009

    Time: 12:16:04 PM

    User: N/A

    Computer: MYNODE1

    Description:

    The MS DTC service was started before the cluster service on this

    cluster node. This service will now be stopped. If MSDTC has been

    configured to run on a cluster, the MS DTC resource will be brought

    online by the cluster service when it starts. Error Specifics:

    d:\nt\com\complus\dtc\shared\mtxclu\mtxclusetuphelper.cpp:932, Pid:

    912, CmdLine: "C:\WINDOWS\cluster\resrcmon.exe" -e 1184 -m 900 -p 224

Page 1 of 2 (20 items) 12