[GE users] manage NFS resources
reuti at staff.uni-marburg.de
Wed Sep 9 14:42:07 BST 2009
Am 09.09.2009 um 15:30 schrieb murple:
>> Do you know inside the job, when the reading of the file finished?
>> There is the feature in SGE to attach meta-data to each job, called
>> job-context. This way you could submit the job with -ac STEP=READING
>> and set it during execution to -sc STEP=EXEC. Having a load sensor on
>> the qmaster or file-server machine could a) build a list of running
>> jobs with 'qstat -s r -u "*" -xml', then b) use a loop across these
>> jobnumbers and count the enrtries 'conext:' in 'qstst -j <jobid>'
>> which read STEP=READING. The load sensor could then be used to
>> trigger a load threshold, when already one-two-three-whatever jobs
>> are in the reading phase.
> Would that also stop other jobs? Because I want other jobs, which
> do not
> use this fileserver still to be able to run.
Depends: one the one hand: yes, the complete queue wouldn't get any
new jobs scheduled. But there is nothing stopping you to define a
dedicated queue for these kind of jobs and only this queue will have
the load_threshold defined. Then other jobs are unaffected. Just make
sure not to oversuibscribe the nodes by defining the maximum number
of slots combined across all queues on a node either in the exechost
definition (qconf -me ...), or in a RQS (resource quota set). You
could use a forced boolean complex to make sure that only your NFS
jobs are going to this special queue.
To unsubscribe from this discussion, e-mail: [users-unsubscribe at gridengine.sunsource.net].
More information about the gridengine-users