Is there any special SCSI command needed to support openVMS Cluster?

[albert000]

As we know, when storage vendors need to support OpenVMS, they need to support SCSI command:A3 or EC to generate the disk letter.

If we want to support OpenVMS cluster, is there any special SCSI command needed to be support?

[Bob Blunt]

Perhaps there's a misunderstanding about what a "SCSI cluster" really is.  There are NO clustering protocols or communication that are sent over any SCSI interface.  ZERO.  All clustering communication (since all the true "hardware" cluster interconnects like CI, DSSI, Memory Channel are not available for Integrity) goes over the "network" (some combination of some flavor of Ethernet ).  IF I recall correctly from conversations with the folk who got it working originally the main changes were to get the SCSI device drivers working correctly with the concept of "seeing" multiple "initiators" on the same SCSI bus.  There were also other challenges with setting up ways to configure systems so device naming could "prevent" duplicate device names.  This was also only possible, in the days of Alpha, with very specific SCSI adapters or HBAs.  There were only a few controllers that could be on the same SCSI bus.  I'm not sure IF or how this works for Integrity-based systems and SCSI interfaces but there would still be similar requirements to get it setup and working properly.

bob

[albert000]

Dear Bob: Thanks for your reply. I have already configured a OpenVMS cluster. However I have met the following problem: The cluster contains 2-nodes(HWNOD1 and HWNOD2). HWNOD1 use the SAN Storage LUN as its system disk. HWNOD2 use local disk as its system disk. This cluster also contains a quorum disk. However when I shutdown one node, the other node will hangs. It can't response to any commands until the abnormal node recovers. Could you tell me how to solve it? Thanks very much. BR tong

[albert000]

Dear Bob:

It seems the forum has some problem with IE. I change to use firefox.

Thanks for your reply.

I have already configured a OpenVMS cluster.

However I have met the following problem:

The cluster contains 2-nodes(HWNOD1 and HWNOD2).

HWNOD1 use the SAN Storage LUN as its system disk.

HWNOD2 use local disk as its system disk.This cluster also contains a quorum disk.

However when I shutdown one node, the other node will hangs. It can't response to any commands until the abnormal node recovers.

Could you tell me how to solve it?

Thanks very much.

BR

tong

[Mike Kier]

What does each system see as the value for QUORUM?  What is the value of VOTES for each system?

Is the quorum disk visible via a direct path (i.e., fiber channel) from BOTH systems?  Either system MUST be able to reach the quorum disk in the event of the failure of the other or you will not have enough votes for quorum.

Are there any quorum disk related messages in OPERATOR.LOG?

Your system should have EXPECTED_VOTES = 3 and a QUORUM of 2 with each node having a VOTE of 1, unless there is some compelling reason otherwise.

[Steven Schweda]

   Perhaps it would be better not to have three threads running for the
same problem.  Certainly, I'd be less confused.  I vote for this one:

      /t5/x/x/m-p/6899759

Along with this one, I'd avoid:

      /t5/x/x/m-p/6899745

[albert000]

Dear

      Thanks for your reply.

       I will use the following links to trace the problem:

      /t5/System-Management/cluster-node-hangs-when-another-node-shutdown/m-p/6899759

      Thanks for all your help.

BR

TONG