[Info-vax] Quorum disk and last transition time

[Stephen Hoffman]

On 2015-01-30 13:22:54 +0000, Joe said:

> On Wednesday, January 28, 2015 at 5:46:52 PM UTC+1, Stephen Hoffman wrote:
> 
>> Any user-mode software and
>> any of the lower-priority kernel-mode VMS code is not going to be
>> stalled,

Correction: user-mode and lower-priority kernel mode code *is* going to 
be stalled; during a quorum hang.

>> I'd not be interested in the cluster state transitions, and would be
>> far more centrally interested in what's up with that quorum disk and
>> whatever bus is involved.
> 
>   Yes, we already identified the issue and working with the provider of 
> storage. This is happening when the disk is served from a different 
> site.

That implies the cluster quorum disk polling timers might be a little 
too tight, or that there are possibly communications errors and/or 
excessive communications latency.

Though probably not applicable here, it is possible to RAID the quorum 
disk locally to avoid outages secondary to hardware errors, but I'd be 
somewhat skeptical about that working right across sites, or across 
controllers.

>   Just got a little curious about how VMS clustering will work in this 
> situation.
>   Thanks, I got the answers :)

Additional Reading:
OpenVMS Tips: Cluster VOTES, EXPECTED_VOTES, and Quorum: 
<http://labs.hoffmanlabs.com/node/153>
Low-End Cluster VOTES, EXPECTED_VOTES <http://labs.hoffmanlabs.com/node/569>
Some of the related parameters: QDSKINTERVAL, RECNXINTERVAL

-- 
Pure Personal Opinion | HoffmanLabs LLC