Quorum out of space

[Jeff Strong]

I have an MSA1000 with two DL380G4s attached in a active/passive cluster arrangement. The quorum is out of space, cluster services won't start, hence I cannot access any of the logical drives on the MSA1000.

Does anyone know how I can accomplish this given the circumstances above? The controller on the MSA1000 has an RJ-45 connection, but I haven't the foggiest (a) how to connect, let alone address the storage.

I am hard down and at a full work stoppage until I can clear some space on that quorum.

Any help would be greatly appreciated.

Thanks!!!!

[Michael Steele_2]

I'm confused or at least I think you're confused about the error message.

Give me some basic information about the O/S, the HW and most of all the LVM data. What utility are you using to make your logical volumes? What server model?

If there is a primary and alternate boot disk in your O/S volume group then you can boot w/out quorum.

[Jeff Strong]

The event log was yielding quorum out of space errors. My guess was that some thing had run it out of space (the physical disk space, not the quorum itself). My issue is trying to get access to my drives on the MSA1000 when I am in the OS so I can actually see what is on those drives as if they're locally attached. Since cluster svc won't come up, I cannot see the drives.

Does that help? Maybe I need more caffeine.

[Víctor Cespón]

The cluster will be using a small logical unit on the MSA as quorum disk. This is the drive that has no free space. OK, you cannot access it from the cluster nodes. I'm no expert in clusters, but I think that you can stop clustering service and access it from a single node like any other LUN. Or you can give access to it to another server and from that server check what files are there.

[Michael Steele_2]

OK - done some reading. Here's a good link about setting up a quorum disk in a MSA1000. Note the comment about accessing the MSA1000 via a serial port. You'll need a laptop or PC for this and then go through hyperterminal which can be found under Programs > Accessories > Communications > hyperterminal. This at least will give you access from the command line of the disk array but no GUI window. For that you need ACU. See this link

http://forums11.itrc.hp.com/service/forums/questionanswer.do?threadId=886338

[Uwe Zessin]

You must fix the problem from Windows.
Even if the quorum disk is a single logical disk on the MSA and you were able to grow it, it would not help, because the MSA does not (and can not) grow the partition table and the file system.

[Kevin Carter_3]

Jeff, the RJ-45 connection on the MSA1000 is for a special serial cable available from HP.

You could create an ACL for the quorum disk that allows another system to access the disk, a system that is not part of the cluster. Then free up some space on the quorum disk from that system and restart the cluster.

Just a thought.

[Tom O'Toole]

You never provided any of the info that was asked for, everyone had to guess the OS was windows. Yeesh.

[Daniel Bernardo]

what's not use te ACU to extend te quorum disk in the msa1000? if you extend the disk you can extend te volumen and your problem are finish.

But the quorum is smaller, have you any other data in the quorum disk??

Pd: sorry for my englihs ;)

[Jeff Strong]

Ok... serial cable a must-get regardless, just for emergency's sake.

I got through the problem within Windows 2KAS and here's how:

An orphaned _vxfiVspCacheFile_0.VSP Symantec NetBackup file. That blew out the space on the quorum, prevented cluster services from coming online, etc, etc. I was able to go to device manager, show hidden devices... then select cluster disk driver, change the start up type to Disabled, OK, reboot. When the node came back online, the storage on the array looked directly attached and I could check data, see what file(s) were crushing the space on my quorum. Once I was able to delete the NBU temp files, I reset Cluster Disk Driver to "System" and rebooted again. My cluster services returned to online status and the cluster came up without incident.

I do, however, need to procure that cable and a guide as to how to use the CLI in the event I cannot get around this. Thank you all for your input. It has been very helpful and is continuing to be.

Jeff