[GE users] Housekeeping

Johnny Layne laynejg at vcu.edu
Tue Jul 24 16:05:48 BST 2007


    [ 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. ]

hye,
    I'm not sure I fully understand what you need to do, but recently I 
wrote a small Ruby script to help me find runaways around our clusters.  
It uses rsh to open a "ps -efwww | grep $TARGET" process on the nodes 
where TARGET is something/someone I want to search for in the output of 
the ps command.  Naturally it takes a minute or so to run & I suppose 
clogs up traffic.  I wanted something to help me track down loose ends 
while I tweaked a tight integration with mpich2 I was working on and 
this did the job; as a bonus I see that I can keep a watch on users as 
well, which is nice as some insist on trying to get around using the SGE. 

    I've tailored this a lot for our particular uses; a parallelized 
version would be awesome!  This just reports processes, if I see 
something interesting in the output I visit the nodes to see what's 
going on.

    If you don't need anything that heavy handed, or don't have the time 
to work on it, then maybe you should check out pdsh, it's really handy 
too. 

    Anyway perhaps a small wrapper script like this to print this output 
in a meaningful fashion would help you.  Good luck!
    johnny

---------------------------------------------------------------------
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