[GE users] Fair Share Policy question

Yogesh Bhanu yogesh at gsf.de
Tue Nov 29 10:54:18 GMT 2005

    [ 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 ,
      I 'm not sure what I'm seeing is right .
I have an SGE6u6 running on Opteron and Pentium cluster .
I have implemented a very simple setup of two cluster wide queues each 
on both architecture .
I 'm not sure if this is a bug or a feature or something wrong they way 
I have setup things up.

I have some users who are competing for cluster resources. I thought 
SGE6 uses fair share system .

In my case what happens is
Say if a user A submits 100 of jobs and the user B submits say 10 jobs .
Then all jobs from A have to finish before jobs from B start to run.
Which is classical case of FIFO which is fine.

Now the anomaly .

I discovered today that when user B had submitted jobs in the morning ..
these jobs  started to run even before the jobs submitted from user A 
yesterday .

Also If I decide to enable  ticket based policy to enable fair share.
How can I do define a default ticket value for all the users in the system.

Any pointers will be of great help


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