[GE users] Question on accessing qmaster spool

Dana Brand dbrand at bcgsc.ca
Tue Jun 3 23:06:46 BST 2008


    [ The following text is in the "ISO-8859-1" character set. ]
    [ Your display is set for the "ISO-8859-10" character set.  ]
    [ Some special characters may be displayed incorrectly. ]

Hi all,

Main question:

We've had an issue with the (NFS shared) qmaster's spooling partition 
running out of space today. ( We use this setup as we rely on 
shadowing).  We want to move the partition to a different filer and yet 
not interrupt the cluster's availability. Here's the question (forgive 
the newbness): is the qmaster spool accessed actively by the head nodes 
only (the ones running sge_master and sge_schedd), or do the execution 
nodes have anything to do with it?

Another question on the side, less important. After we have made more 
room for the spool on that partition, the qmaster's messages file 
started getting flooded every 15 seconds (sched. interval) and never 
stopped since with messages like:
06/03/2008 15:02:48|qmaster|apollo|W|scheduler send a order for a 
changed user/project "Drug" (version: old 49391) new 49392
where the above was a project that wanted to save some information 
during the outage:
06/03/2008 06:01:18|qmaster|apollo|E|can't open file projects/.Drug: No 
space left on device
Any ideas on how to stop these?
Thanks a lot for your support,
Dana

-- 
Dana Brand	
Computer Systems Specialist
Genome Sciences Centre
Tel: (604) 707-5900, ext. 5424


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe at gridengine.sunsource.net
For additional commands, e-mail: users-help at gridengine.sunsource.net




More information about the gridengine-users mailing list