• Status: Solved
  • Priority: Medium
  • Security: Public
  • Views: 599
  • Last Modified:

2003 Enterprise and 2003 Enterprise R2 Cluster Nodes

I am experiencing issues joining a 2003 Enterprise R2 cluster node to our production cluster.  I have 2 nodes with HBA cards directly attached directly to a Nexsan Sataboy unit with 1 storage controller and two fiber ports.  It is an Active/Passive File cluster.  Node 1 connects to fiber port 1 on Nexsan.  Node 2 connects to fiber port 2.  Both existing production nodes are running Server 2003 Enterprise SP2.  My goal is to replace/ upgrade each node with a new server (2003 ENT R2 SP2) one  at a time:

1 - Evict old node running 2003 Enterprise SP2
2-  Join new Node running 2003 Enterprise R2 SP2

I can successfully join new 2003 Enterprise R2 SP2 node to cluster.  However, when attempting to move group or fail over to new cluster node, It reverts right back to original cluster node.  The system event log displays Error ID 1034:

The disk associated with cluster disk resource 'SATABOY Q:F:' could not be found. The expected signature of the disk was E611A4A1. If the disk was removed from the server cluster, the resource should be deleted. If the disk was replaced, the resource must be deleted and created again in order to bring the disk online. If the disk has not been removed or replaced, it may be inaccessible at this time because it is reserved by another server cluster node.

I am using exact model of HBA card on new server with updated Driver, FW, and card management utility on new node.  I also tried the same scenario with card taken from production node using same firmware, driver and card management version used in production.  I still get the same error.  

I can revert back to original cluster node and everything works perfectly.

I have tried joining new node to cluster (node 2, 2003 Ent R2 SP2), rebooting node 1 and attempting failover again with same results.  As soon as I revert back to original node 2 with 2003 Ent SP2, everything starts working fine.

Nexsan is set to map ports to proper LUN with R/W access.

I did notice a difference in HBA card software between old and new node.  Old node sees Nexsan storage device as SCSI3 device.  New node with 2003 R2 sees Nexsan as SCSI2 device.  HBA drivers and software act as transport only and OS Device negotiation (SCSI2/SCSI3) is between OS and Nexsan.

I have Nexsan looking into issue currently.

No drives have been removed or replaced.

Are there any known issues with joining 2003 Ent R2 to cluster with other nodes running 2003 ENT SP2?

Is there a difference in the ENT 2003 SP2 vs ENT 2003 R2 SP2 OS in how they view SCSI devices?



0
marktgeorge
Asked:
marktgeorge
  • 5
1 Solution
 
E037600Commented:
MS Cluster would see a difference between ENT 2003 SP2 and ENT 2003 R2 SP2.  Refer to this document, it is for upgrading a 2-node 2003 cluster to 2008, but I suspect your steps would be the same.

http://technet.microsoft.com/en-us/library/cc753627(WS.10).aspx

0
 
marktgeorgeAuthor Commented:
Where is it documented that it would see a difference or that they will not work together?
0
 
marktgeorgeAuthor Commented:
Got it working.

There were two issues:

1)  KB932465 - Nexsan unit was set up as GPT volume.  Had to download and install hot fix 919117 to add support for GUID partition table (GPT) volumes that are larger than 2 terabytes on a Windows Server 2003-based server cluster.  Installed this hot fix on new R2 node.

2) I was using the same model of HBA card with new driver, firmware, and card management software.  Nexsan firmware is outdated and was not properly communicating with newer software/drivers/firmware on Emulex HBA card.  Removed HBA card from old server and installed in new server.  Reverted to working older production drivers and card management software.  After reboot, new node ran like a champ.  Fail over and group moves were successful.  No more issues.

I will be replacing the second node next week and will use the same attempted card with the older drivers and card management software.  This card has the latest firmware on it but I suspect that the issue is with the driver.  The version of card software management utility is dependent upon the driver and not the firmware.  If my initial attempts to install succeed, this will confirm it.  Otherwise, I will take the existing card out of the old node and place it in the new node and continue.

Once I isolate to see if the card communication issue is related to the drivers and software or drivers software, and firmware I will update this thread.

Thanks for responding E037600.
0
Industry Leaders: We Want Your Opinion!

We value your feedback.

Take our survey and automatically be enter to win anyone of the following:
Yeti Cooler, Amazon eGift Card, and Movie eGift Card!

 
marktgeorgeAuthor Commented:
As an addition to Issue 1, you can verify that your Nexsan is set up as a GPT volume by opening a command prompt and using the diskpart utility.

log into node with control of cluster physical disk resource (Shared Storage)
 
type in:
Diskpart
list disk

Look for shared storage disk number which you can confirm in computer management under Disk management.

If it is GPT, you will see an * under the GPT category on the right hand side.
0
 
marktgeorgeAuthor Commented:
HBA card with new firmware with older drivers and software worked.  It should work with newer drivers and software if I update the Nexsan unit firmware the HBA card is directly attached to, but that is for another day.


0
 
marktgeorgeAuthor Commented:
After troubleshooting, the issue was resolved by a hotfix for GPT volumes and software selection for card used to communicate with Nexsan device.
0

Featured Post

Hire Technology Freelancers with Gigs

Work with freelancers specializing in everything from database administration to programming, who have proven themselves as experts in their field. Hire the best, collaborate easily, pay securely, and get projects done right.

  • 5
Tackle projects and never again get stuck behind a technical roadblock.
Join Now