On Thu, Dec 10, 2009 at 08:34:12AM -0800, Jim Kusznir alleged: > Ok, I increased the logging level and tried the various definations of > RMCFG suggested by others; no chage or helpful results. > > I think this is the crux of the problem (loglevel 5): > > 12/10 08:20:49 MRMWorkloadQuery() > 12/10 08:20:49 MPBSWorkloadQuery(ISP-CURRAN.ISP.WSU.EDU,JCount,SC) > 12/10 08:20:49 INFO: queue is empty > 12/10 08:20:49 INFO: 0 PBS jobs detected on RM ISP-CURRAN.ISP.WSU.EDU > 12/10 08:20:49 WARNING: no workload detected > > Just a few lines above that, it does: > 12/10 08:20:49 MRMClusterQuery() > 12/10 08:20:49 MPBSClusterQuery(ISP-CURRAN.ISP.WSU.EDU,RCount,SC) > 12/10 08:20:49 __MPBSGetNodeState(Name,State,PNode) > 12/10 08:20:49 INFO: PBS node isp-curran set to state Idle (free) > > And correctly detects my resources and their states. > > So maui is clearly able to talk to torque, its just not able to load > the list of jobs. It does load the compute resources just fine. Does > maui use a different method/connection to load jobs from the scheduler > (i.e., a different network port to torque, etc)?
It's a permissions issue. You need to set the maui user as a manager in pbs_server. 'set server managers += m...@isp-curran' will probably do what you want. In the pbs_server log, you probably have a message about m...@somewhere connecting. It's not an error message. Use the same u...@host string it is printing. I know you already have it in the list with the fqdn, but pbs_server only cares about the results of the reserve lookup. It doesn't know your domain name and doesn't know that isp-curran == isp-curran.isp.wsu.edu. -- Garrick Staples, GNU/Linux HPCC SysAdmin University of Southern California Life is Good!
pgpVJopRdwpRr.pgp
Description: PGP signature
_______________________________________________ mauiusers mailing list [email protected] http://www.supercluster.org/mailman/listinfo/mauiusers
