SQL Server Performance Forum – Threads Archive

Cluster Service using 2-node

During installation of SQL Server 2K on a 2-node cluster with Windows Server 2003 , do I need to have both Cluster Service running on both primary and secondary nodes? I am trying to install the first instance of SQL server on my primary node but I kept on receiving below message when ask to "Enter a user name and password that is a valid administrator account for all nodes in the cluster system" even though I have already setup the necessary admin account for both nodes:
"The specified account cannot be validated to have administrator rights on NODE2.
An error occured: (1203)
No network provider accepted the given network path." Thanks,
Jon M

Yes, when you install SQL Server 2000 onto a cluster, both nodes have to be up and running, as the setup program installs to both nodes at the same time. —————————–
Brad M. McGehee, MVP
Webmaster
SQL-Server-Performance.Com

Thanks Brad. I think I found the reason why am I getting below message. I enabled the File and Printer Sharing for Microsoft Networks on both of my nodes and now I have completed the first stage of my SQL cluster. I will install the SQL server SP3a now… Brad, My objective is to come up with multiple-instance sql server (active/active). Do I have to install the SP3a first on my primary node and then install another instance on the secondary followed by the SP3a? Thanks. Wooohooo!!! Jon M [<img src=’/community/emoticons/emotion-2.gif’ alt=’

‘ />]

SQL Server service packs have to be installed separately on each node. In theory, the setup program should install it on both nodes, but it does not appear to work. I have always had to do them separately. —————————–
Brad M. McGehee, MVP
Webmaster
SQL-Server-Performance.Com

>>>
SQL Server service packs have to be installed separately on each node. In theory, the setup program should install it on both nodes, but it does not appear to work. I have always had to do them separately.
>>> really? I’ve never had any issues with sp3 on the w2k3 cluster… I had to install SP3 for each sql instance but only on the node currently running the instance, not on all nodes…? Twan

In my case, these were Windows 2000 clusters. I don’t know if that makes any difference. —————————–
Brad M. McGehee, MVP
Webmaster
SQL-Server-Performance.Com

After installing the latest sql server service pack in my primary node, i don’t see my shared disks anymore. [<img src=’/community/emoticons/emotion-6.gif’ alt=’

‘ />] Any idea? Thanks, Jon M

Did the instance failover to the second node and not return back to the first node? Therefore the disks would be visible from node 2 but not from node 1? Cheers
Twan

Twan, I have not reached that point. Anyway, since I don’t see my shared disk, I decided to reboot/reset everything including the shared disk. Afterwards, my cluster went normal again with the latest sql server SP and I can see the shared disks. I proceeded and performed below test written by Brad:
http://www.sql-server-performance.com/sqlserver2000_clustering_install_part4.asp I passed Test #s 1 & 2 but failed in Test #3. I physically turned off the primary node and saw cluster groups failed over to the secondary node. But when I did the same for the secondary node, my cluster got messed up again. Checking the event viewer for my primary node, I saw a lot of messages such as below:
-Cluster service is requesting a bus reset for device DeviceClusDisk0.
-The device, DeviceHarddisk3, is not ready for access yet. I think my problem is more on the OS cluster, HW setup…I will investigate further. Jon M

I am now testing my 2-node sql active/passive cluster. After breaking the primary node’s public network, all failed-over to my secondary node. But after plugging back in the network cable of my primary node, all automatically failbacked from secondary to primary except for the SQL Server, SQL Server Agent, and SQL Server Fulltext which remained on the secondary node. What am I doing wrong? Thanks,
Jon M

How can they be left on the secondary node? Either the entire group fail over or nothing fails over. In the cluster you should have at least:
Cluster Group
SQL Group (named as your instance) In the SQL Group you have:
Disks
SQL IP
SQL Name
SQL Server
SQL Agent
SQL Fulltext In the Cluster group you have:
Cluster IP
Cluster Name
Quorum disk
MSDTC (unless it has it’s own group) Check the Windows Event Logs on both nodes for error messages.

Argyle, Yes I have the same setup. After rebooting the primary node, the SQL group failed back from secondary to primary. Is this normal? Checking the event viewer…
From the application event viewer, I saw a warning:
"Replication of license information failed because the License Logging Service on server \myserver could not be contacted." From the system event viewer:
1. Error: "Cluster service is requesting a bus reset for device DeviceClusDisk7Part0."
2. Warning: "The driver for device DeviceScsiql23001 performed a bus reset upon request."
3. Error: "The device, DeviceHarddisk7, is not ready for access yet." Jon M

It depends on your group setting. If you right click the group in cluadmin and check the Failback tab and you have it set to Prevent failback then it’s not normal. If you have it set to Allow failback then it’s normal.

]]>