quorum disk : ALPHA OPEN VMS

[pavithra123]

Hi All, I am seeing the following errors in the operator.log, For some weird reason the diagnose utility ( DEC event) has not been installed, So I cannot analyse the errorlog.

%%%%%%%%%%% OPCOM 27-JUL-2004 09:43:20.51 %%%%%%%%%%%
09:43:20.51 Node SMST21 (csid 00010008) failed to read quorum disk

%%%%%%%%%%% OPCOM 27-JUL-2004 09:43:20.51 %%%%%%%%%%%
09:43:20.51 Node SMST21 (csid 00010008) lost "connection" to quorum disk

%%%%%%%%%%% OPCOM 27-JUL-2004 09:43:22.20 %%%%%%%%%%%
09:43:22.20 Node SMST21 (csid 00010008) re-established "connection" to quorum disk.

I am runnig the OPEN VMS (7.3.1 ) on alpha DS20 machines. Its a dual node running on cluster.

Any help is greatly appreciated.

Prem.

 

[mjpearson]

I know enough to get into trouble but getting out is a different story so......take my observations with caution:

Observation:

The "read" and "connection" problems are back-to-back. The "reconnection" happens less than 2 seconds later. I wonder if the disk had a hiccup or there might have been a power dip that caused the disk to slow down?


mike

 

[HughSamson]

Hi Prem..
Could you supply a little more info regarding the configuration:
What storage are you using - SAN, scsi BA shelves etc that the quorum disk resides on.
The position of the quorum disk on the i/o bus (ideally it should be the first disk on the i/o bus for the speed).
Is there anything else on the disk (sharing with application data maybe?).
You should keep in mind, if it isn't already configured this way, that the quorum disk requires a physical i/o when connecting to the disk, it cannot be cached - so it must be fast access - a couple of seconds delay in the i/o will cause these messages.
There are a couple of config parameters in modparams you can tweak to set the timeout higher, which may suppress these messages - but don't set them too high, or you are asking for trouble...
Cheers,
Hugh.