IBM®
Skip to main content
    Country/region [select]      Terms of use
 
 
    
     Home      Products      Services & solutions      Support & downloads      My account     
 
developerworks > My developerWorks >  Dashboard > HPC Central Wiki > ... > General Parallel File System (GPFS) > Quorum
developerWorks
Log In   View a printable version of the current page.
Overview Connect Spaces Forums Wikis
Quorum
Added by ScottGPFS, last edited by ScottGPFS on Nov 20, 2008  (view change)
Labels: 
(None)

This page provides additional information on how quorum works in GPFS.

Topics

  • If there are two nodes using quorum with tiebreaker disks and network connection between them is lost, who wins?

If there are two nodes using quorum with tiebreaker disks and network connection between them is lost, who wins?

If the Cluster Manager node still has access to a majority of the tiebreaker disks, then the nodes that can still communicate with the Cluster Manager node will form the surviving group.

In GPFS release 3.2 you can move the Cluster Manager using

mmchmgr -c $nodename

One anomaly of this is that the Cluster Manager may be the only remaining node if its network gets disconnected. To prevent this you can define a minimum number of nodes:

mmchconfig minQuorumNodes=2

This requires that the Cluster Manager have a co-conspirator.


 
    About IBM Privacy Contact