[GE issues] node fill-up issue
gascan211
gascan211 at gmail.com
Mon Dec 28 21:50:40 GMT 2009
Hi,
I want the default behavior of sge scheduler to use the least used nodes first and then fill up the available slots on a node.
queue_sort_method is set to load and np_load_avg=0.50 (normal usage)
But nodes are being filled up instead of jobs being scheduled on the least used nodes. It happens only when jobs are submitted in batches. If jobs are submitted at various intervals, they are being dispatched to different nodes. I.e. submit a job, wait until it is running, submit the next job, and so on.
We haven't seen this issue in SGE6.0u9. Is this a bug in 6.2u1
Thanks,
Kumar
Scheduler configuration:
algorithm default
schedule_interval 0:0:5
maxujobs 0
queue_sort_method load
job_load_adjustments np_load_avg=0.5
load_adjustment_decay_time 0:7:30
load_formula slots
schedd_job_info false
flush_submit_sec 0
flush_finish_sec 0
params none
reprioritize_interval 0:0:0
halftime 168
usage_weight_list cpu=1.000000,mem=0.000000,io=0.000000
compensation_factor 5.000000
weight_user 0.250000
weight_project 0.250000
weight_department 0.250000
weight_job 0.250000
weight_tickets_functional 0
weight_tickets_share 0
share_override_tickets TRUE
share_functional_shares TRUE
max_functional_jobs_to_schedule 200
report_pjob_tickets TRUE
max_pending_tasks_per_job 50
halflife_decay_list none
policy_hierarchy OFS
weight_ticket 0.500000
weight_waiting_time 0.278000
weight_deadline 3600000.000000
weight_urgency 0.500000
weight_priority 0.000000
max_reservation 0
default_duration INFINITY
------------------------------------------------------
http://gridengine.sunsource.net/ds/viewMessage.do?dsForumId=36&dsMessageId=235318
To unsubscribe from this discussion, e-mail: [issues-unsubscribe at gridengine.sunsource.net].
More information about the gridengine-users
mailing list