Physical Disk Resource will not go online 1

[stelljohann]

This is a brand new installation on new equipment. I'm having an issue with the physical disk resource. Here's the scenario:
2 Dell PowerEdge 2850 servers running Windows 2003 Ent.
1 Dell PowerVault 220S SCSI Array w/4 146GB 15k drives.
The drive array is in cluster mode connected to both servers.

Server 1 is powered up, Server 2 is off. I can connect to the drive array without any problems. I can see the logical drives that I created without any problems and can format them and assign drive letters. I can write files to the newly created drives just fine. I then go into cluster Administrator and create a new cluster using the wizard.

The cluster creates fine until it tries to bring the physical disk resources online (Q: and R: - both on the SCSI array). The errors it gives in the cluster log are:

Resource 'Disk Q:' failed. Check Event Log. Dependent resources might fail too. (hr=0x80004005, {58C977AB-7B70-4D9E-988C-F521DA5F3829}, {FB907A0C-E4E5-413C-A8E1-DE148C04EE0F}, 1, 1, 1)

Resource 'Disk R:' failed. Check Event Log. Dependent resources might fail too. (hr=0x80004005, {58C977AB-7B70-4D9E-988C-F521DA5F3829}, {55C6240A-94FA-4E75-933D-2728C452C33C}, 1, 1, 1)



Thanks,
stelljohann
--The number of times an uninteresting thing happens is an interesting thing.--

 

[xmsre]

what version of openmangle/array manager?

 

[stelljohann]

Dell OpenManage Server Administrator Version 5.0.0

Thanks,
stelljohann
--The number of times an uninteresting thing happens is an interesting thing.--

 

[xmsre]

If you right click the resource and click advanced, what is the "pending timeout" value set at?

 

[stelljohann]

180 seconds

Thanks,
stelljohann
--The number of times an uninteresting thing happens is an interesting thing.--

 

[xmsre]

80004005 - E_FAIL. At first I thought about issues with the arry manager component of openmange prior to 5.0, but you're using 5.0. Then I thought perhaps the pending timeout was too short for the disk to come on line, but 180 seconds should be more than adequate. Going back and rereading the description, I have a question:

How is the SCSI bus terminated? If you are not using "Y" cables and external termination, then the termination is at the host. If this is the case, then having node B powered off means your SCSI bus is not terminated.

 

[stelljohann]

I found the solution here:

http://forums.us.dell.com/supportforums/board/message?board.id=pes_hardrive&message.id=21625

Basically, I created 2 RAID 0 drives, then went into the 'object' option in the PERC adapter configuration and set the cluster option to 'enabled'. Once the cluster option is enabled, then the SCSI ID field can be edited. This only needed to be changed on the second node. After that, everything was fine. Thanks for the suggestions. Your SCSI termination question got me to thinking, and eventually lead me down the right path.

Thanks,
stelljohann
--The number of times an uninteresting thing happens is an interesting thing.--