HP OpenVMS Systems

Quorum Disk Connection Errors?

» close window

The Question is:

 
I have a alphaserver ds10 running v 7.3.-1.  I intend to add another node to
 this computer so I ran cluster_config to prepare for this.
Cluster_config added to modparams.dat:
votes=1
vaxcluster=2
disk_quorum="dka0"
dqskvotes=1
mscp_load=1
mscp_serve_all=1
bootnode="Y"
My problem is that when I am running backup on the system disk (dka0:) then I
 get "failed to write quorum disk" and "lost "connection" to quorum disk".
Do I need to change the values in modparams.dat?
And also what are the best values when I add the other node.
Thank you.
 

The Answer is :

 
  Please see the discussion of VOTES and EXPECTED_VOTES in the
  OpenVMS FAQ.
 
  Also please realize that there is NO use for a quorum disk
  that is not accessable through multiple paths; that is not
  directly accessable to at least two hosts.  You are better
  to assign the votes to the host system, in this case.
 
  Please check to see if the device DKA0: is accessasble.
 
  If the target of the BACKUP directly or indirectly the DKA0:
  disk or I/O bus, please realize that BACKUP can saturate the
  device and this can trigger quorum-related errors as the
  quorum I/O operations become blocked by the torrent of I/O
  and the timers fire.  You can avoid these errors by reducing
  the I/O load to the quorum disk or by moving the quorum disk
  to another (and quieter) spindle.  You can potentially suppress
  the errors by increasing the QDSKINTERVAL setting.
 

  

     

     
answer written or last revised on ( 11-APR-2003 )
     
» close window