I have discovered a problem when running with 
AccountingStorageEnforce=limits,qos.  To demonstrate the error I set 
maxsubmit=4, submitted 4 jobs & did an 'scontrol reconfigure'.  I then did 
an 'scancel' of one of the  jobs & submitted another job.  This results in 
the message 'srun: error: Unable to allocate resources: Job violates 
accounting policy (job submit limit, user's size and/or time limits)'.  I 
had instrumented the logic in acct_policy.c  function 
_adjust_limit_usage() which increments and decrements the submit_jobs 
count.  What I see is that the 'reconfigure' causes the count to be 
incremented a second time for jobs that are active in the system (in this 
case either pending or running).  The submit_jobs count was 8 prior to my 
issuing the 'scancel' command.  The submit_jobs count is being updated as 
a result of the read_config.c function _restore_job_dependencies calling 
acct_policy_add_job_submit().  Are there any suggestions on how to resolve
this problem?

Best Regards,
Bill

Reply via email to