Hello,

I'm using SLURM 2.6.2

I have set a reservation on the full cluster :

scontrol show reservation
ReservationName=root_10 StartTime=2013-11-08T08:00:00
EndTime=2013-11-08T12:00:00 Duration=04:00:00
   Nodes=node[01-56] NodeCnt=56 CoreCnt=896 Features=(null)
PartitionName=(null) Flags=MAINT,IGNORE_JOBS,SPEC_NODES
   Users=root Accounts=(null) Licenses=(null) State=INACTIVE

I'm quite surprised to see that jobs submitted now for a duration of four
days are still scheduled before the beginning of the reservation date. If
the job is really launched at this time it will overlap the reservation
duration. The jobs have been submitted after the reservation was set.

scontrol show job 53951
JobId=53951 Name=myjob
   UserId=user1(11111) GroupId=mygroup(1000)
   Priority=3016 Account=myaccount QOS=normal
   JobState=PENDING Reason=ReqNodeNotAvail Dependency=(null)
   Requeue=1 Restarts=0 BatchFlag=1 ExitCode=0:0
   RunTime=00:00:00 TimeLimit=3-23:59:00 TimeMin=N/A
   SubmitTime=2013-11-05T13:23:31 EligibleTime=2013-11-05T13:23:31
   StartTime=2013-11-06T14:18:58 EndTime=Unknown
   PreemptTime=None SuspendTime=None SecsPreSuspend=0
   Partition=parallel AllocNode:Sid=master:178324
   ReqNodeList=(null) ExcNodeList=(null)
   NodeList=(null)
   NumNodes=1-1 NumCPUs=1 CPUs/Task=1 ReqS:C:T=*:*:*
   MinCPUsNode=1 MinMemoryNode=0 MinTmpDiskNode=0
   Features=(null) Gres=(null) Reservation=(null)
   Shared=0 Contiguous=0 Licenses=(null) Network=(null)
   Command=mycommand
   WorkDir=myworkdir

I'm using PriorityType=priority/multifactor

Any idea?

Reply via email to