[Mauiusers] jobs not FIFO scheduled

Marc Noguera marc at klingon.uab.es
Fri Jan 30 02:25:28 MST 2009


Hello all,
I am still this strange scheduling problem. However, the problem has
turned to jobs not running even if resources are available. Note that if
maui is restarted these jobs are started immediately.
For instance, "qstat -na1 qserq1" (which is  a name of a queue/class)
results in:
/----------
quanta.uab.es:
                                                                        
Req'd  Req'd   Elap
Job ID               Username Queue    Jobname          SessID NDS   TSK
Memory Time  S Time
-------------------- -------- -------- ---------------- ------ ----- ---
------ ----- - -----
42144.quanta.uab     vicenc   qserq1   modelcis-MeCN2pq  10453     1 
--    --    --  R 90:45   borg59
42958.quanta.uab     vicenc   qserq1   AUUAi-nw-m062x.n    --      1 
--    --    --  Q   --     --
42960.quanta.uab     marc     qserq1   gli_g03.dat         --      1 
--    --    --  Q   --     --
-----------/

Jobs 42958 and 42960 are not running but there are procs available.
"Checkjob" results in:

/---------------
checking job 42958

State: Idle
Creds:  user:vicenc  group:serq  class:qserq1  qos:DEFAULT
WallTime: 00:00:00 of 99:23:59:59
SubmitTime: Fri Jan 30 09:56:07
  (Time Queued  Total: 00:20:32  Eligible: 00:20:32)

Total Tasks: 1

Req[0]  TaskCount: 1  Partition: ALL
Network: [NONE]  Memory >= 0  Disk >= 0  Swap >= 0
Opsys: [NONE]  Arch: [NONE]  Features: [opteron][qserq1]
NodeCount: 1


IWD: [NONE]  Executable:  [NONE]
Bypass: 0  StartCount: 0
PartitionMask: [ALL]
PE:  1.00  StartPriority:  20
job can run in partition DEFAULT (3 procs available.  1 procs required)
-----------------------------/
 
So, maui sees that 3 procs are available, but does not proceed to run
the job. Output is the same for the other job.
This problem occurs on other queues as well. There are some other jobs
that have been in the Q state for quite a long but in other queues, not
in this one or some other that experience this problem.
diagnose -p command results in:

/---------------
diagnosing job priority information (partition: ALL)

Job                    PRIORITY*   Serv(QTime)
             Weights   --------       1(    1)

42924                      1032   100.0(1032.)
42935                       959   100.0(959.4)
42936                       959   100.0(959.3)
42958                        22   100.0( 22.1)
42959                        21   100.0( 20.8)
42960                         5   100.0(  4.5)
42961                         4   100.0(  4.2)

Percent Contribution   --------   100.0(100.0)

* indicates system prio set on job
----------------/

So there are jobs  with higher priorities, but they are in other queues.

Following your advice, my maui.cfg is:
/----------
SERVERHOST            quanta.uab.es
ADMIN1                mauiuser root
RMCFG[QUANTA.UAB.ES] TYPE=PBS
AMCFG[bank]  TYPE=NONE
RMPOLLINTERVAL        00:00:30
SERVERPORT            42559
SERVERMODE            NORMAL
LOGFILE               maui.log
LOGFILEMAXSIZE        1000000000
LOGLEVEL            2
QUEUETIMEWEIGHT       1
BACKFILLPOLICY  NONE
RESERVATIONPOLICY     CURRENTHIGHEST
NODEALLOCATIONPOLICY  MINRESOURCE
DEFERTIME 0
CLASSCFG[parallel]
CLASSCFG[qdynamics6]
CLASSCFG[qdynamics7]
CLASSCFG[qgetab1]
CLASSCFG[qgetab2]
CLASSCFG[qstruct2]
CLASSCFG[qserq1]
CLASSCFG[qserq2]
CLASSCFG[transmet2]
CLASSCFG[transmet3]
CLASSCFG[transmet4]
CLASSCFG[transmet5]
CLASSCFG[transmet1]
CLASSCFG[qstruct1]
CLASSCFG[nahaste]
CLASSCFG[obelix]
CLASSCFG[joanpauii]
CLASSCFG[soca]
-----------------------------
/
What am I doing wrong? Any help is appreciated.

Thanks in advance

Marc


Gabe Turner escribió:
> On Wed, Jan 28, 2009 at 10:04:01AM +0100, Marc Noguera wrote:
>   
>> Thank you,
>> I made the change on maui.cfg and set BACKFILLPOLICY to NONE. However
>> jobs are not running if there is a previous job in queue in any other
>> queue, not in the same queue
>> Does that make any sense?
>>     
>
> It's possible that happening because you've set RESERVATIONPOLICY to NEVER.
> That means no job will ever have a reservation.  Try unsetting that.
>
> Gabe
>   



More information about the mauiusers mailing list