OpenVms trying to talk to the quorum disk

[Kirk Reindl]

Hello,

I'm new to OpenVMS, I was formerly an HPUX admin.

I've been given this responsibility to administer a two-node cluster running openvms 7.3-1.

Well, I'm noticing the following quorum issues: (I've already logged a call to support, but I thought I'd put something out to the forum.

Here are the messages that are being output to the screen:

%%%%%%%%%%% OPCOM 14-JAN-2004 15:11:57.02 %%%%%%%%%%%
15:11:57.02 Node OSIJX1 (csid 00010003) timed-out operation to quorum disk

SYS_OSIJX1>show dev d
%%%%%%%%%%% OPCOM 14-JAN-2004 15:11:57.02 %%%%%%%%%%%
15:11:57.02 Node OSIJX1 (csid 00010003) lost "connection" to quorum disk

SYS_OSIJX1>show dev d
%%%%%%%%%%% OPCOM 14-JAN-2004 15:12:17.95 %%%%%%%%%%%
15:12:17.95 Node OSIJX1 (csid 00010003) re-established "connection" to quorum disk

Could anyone give me some direction as to what the problem might be?

Thanks
Kirk Reindl

[labadie_1]

check you have all the patches for vms 7.3-1.

Does it happen when the quorum disk has a lot of I/O (backup or anything doing a big I/O activity) ?

Anyway you should install Amds or Availabilty Manager, available at

http://h71000.www7.hp.com/openvms/products/availman/index.html

I remember a 2 node Cluster with a quorum disk where suddenly all hang: somebody offered to power-off and reboot the whole Cluster. A wiser man started Amds, saw that the quorum disk was offline, asked Amds to re-calculate the quorum, and all was fine :-)

[Martin P.J. Zinser]

Hello Kirk,

welcome in VMS Land!

As for yuor problem, how is the quorum disk connected to the two nodes? If you do a show device on the quorum disk on the two nodes, do you see error counters increasing (you need to look on each node separatly).

Greetings, Martin

[Lokesh_2]

Hi,

This error indication connection problem with your quorum disk.
To see which is your quorum disk, use the following command:

$ mc sysgen sho disk_quorum
Parameter Name Current Default Min. Max. Unit Dynamic
-------------- ------- ------- ------- ------- ---- -------
DISK_QUORUM "$1$DUA214 " " " " " "ZZZZ" Ascii
$


HTH,
Best regards,
Lokesh

[Ian Miller. · Accepted solution]

I expect hardware problem talking to q disk or large i/o load delaying the regular polls of the q disk. System parameter QDSKINTERVAL controls how often the system will poll.

As you are new to VMS I suggest you buy a copy of "OpenVMS System Management Guide 2nd Edition" which is an excellent book.

[Mike Naime]

Kirk:

Wealcome to VMS.

Is this happening during a backup of the system (OS) drive? If so, this is a normal occurence during the backup process.


I would guess from looking at your time stamps on the OPCOM messages that your Qdiskinterval = 20. This is a fairly standard value.

Increasing or decreasing this value can have both good and bad consequences. You have to understand what is going on before you change it (If at all)

If this is a SAN based drive, there may be no real reason for the messages. We asked Compaq to investigate these types of messages 3 years ago when we first started seeing them on our SAN fabric. (50 2-node clusters that boot from the SAN) Basically it was explained as part of doing business on the SAN fabric. We now ignore them as noise messages.

Your only real cause for concern would be if this caused your systems to pause because this triggering a cluster transition that made your systems halt for the QDKSINTERVAL time.


Mike Naime

[Kirk Reindl]

Thanks for all responses and suggestions. Unfortunately I'm not able to make a move/change at this moment; I'm buried in work and unfortunately I'm acting in reactive mode. I'll save all responses for when I need them in the future.

Kirk Reindl