[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] jobmanager other than Fork, LSF and PBS



Hi Diego...,

Well thanks for your review...
I have worked a week with condor... and submitted the jobs from condor submission machine to condor host.. and host was able to delegate the jobs to the condor pool and the result is posted back to submission machine... this was kinda C jobs..

well now moving step ahead... now throwing jobs to GT 4...using Condor...
this is my submission file... lot is done to improve it...

$ cat env3test
executable = /bin/ls
universe = grid
grid-type = gt4
#globusscheduler = i2it:8443
JobManager_type = Fork
#universe = globus
grid_resource = gt4 https://192.168.7.131:8443/wsrf/service/ManagedJobFactoryService Fork
output = test_env3.out
log = test_env3.log
error = test_env3.err
notification = error
#requirements = (OpSys == "LINUX")
queue 1

considering ur mail and reading text .. got that Fork is default scheduler...
now when i submit this ...
$ condor_q -ana
-- Submitter:xxxxxxxxxxxxxx : <192.168.7.131:32849> : xxxxxxxxxxxxxxxxxx
 ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD
  78.0   pranavg         8/17 15:38   0+00:00:00 H  0   9.8  ls
  79.0   pranavg         8/17 15:38   0+00:10:59 R  0   9.8  gridftp_wrapper.sh
 
2 jobs; 0 idle, 1 running, 1 held

this 79.0 was greated after 2 minutes of submission of 'env3test' 
well i dont know why.. explain me if u get the answer...

now the problem... well looking at the log file it says so...

000 (078.000.000) 08/17 15:38:10 Job submitted from host: < 192.168.7.131:32849>
...
018 (078.000.000) 08/17 15:38:24 Globus job submission failed!
    Reason: 0 The AXIS engine could not find a target service to invoke!  targetService is null
...
012 (078.000.000 ) 08/17 15:38:28 Job was held.
        Globus error: The AXIS engine could not find a target service to invoke!  targetService is null
        Code 0 Subcode 0
...

could you tell what is the probable problem with this... and the solution as well :-)

Thanx...
Pranav.......

On 8/16/06, Diego Bello <dbello@xxxxxxxxx> wrote:
On 8/16/06, Pranav Aggarwal <pranav.aggwl@xxxxxxxxx> wrote:
> Hi,
>
> looks like i completely dumped myself...
> To begin with...
> I have a machine which has gt4 installed and condor v6.8.0 full install...
> simple jobs running from condor to other condor pool which is managed by a
> condor host manager...
> now i wanna try some globus jobs... so i used a description file as:
>
> $ cat env2test
>
> executable = env-test.pl  *
> globusscheduler = pal-w2p.corp.cdac.in/jobmanager **
> universe = grid
> grid-type = gt4
> #universe = GLOBUS
> output = env2test.out
> log = env2test.log
> Transfer_Executable = TRUE
> queue
>
> * env-test.pl runs well when u fire command...
> $ perl env-test.pl
>
> ** pal-w2p.corp.cdac.in is the ip name with domain.... and its of my
> machine...
>
>
> to tell you more... i dont have any job manager such as FORK, LSF or PBS but
>  condor is handling all the things... i guess... (correct me if i am wrong)
>
> whenever i fire from non-root and non-condor user...
> $ condor_submit env2test
>
> the job submits in idle state...
>
> $ condor_q -ana
>  ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD
> ---
> 045.000:  Run analysis summary.  Of 2 machines,
>       0 are rejected by your job's requirements
>       0 reject your job because of their own requirements
>       0 match but are serving users with a better priority in the pool
>       2 match but reject the job for unknown reasons
>       0 match but will not currently preempt their existing job
>       0 are available to run your job
>
> WARNING: Analysis is only meaningful for Globus universe jobs using
> matchmaking.
> 1 jobs; 1 idle, 0 running, 0 held
>
> could you plz help me what is going wrong... thanx...
>
> Pranav
>

First of all, you should test Condor first, without doing anything
with Globus. Try submitting some simple jobs to Condor viewing the
documentation. The rejecting problems always happen when setting up
Condor.

Second of all, Globus, by default, has the FORK jobmanager, it is the
way globus manages jobs execution. Also, globus has the option to
compile a jobmanager for condor, so you submit a job to
globus_host/jobmanager-condor, globus takes the job, autenticates your
account and then passes the job to Condor.

Regards.
--
Diego Bello Carreño
Estudiante Memorista de Ingeniería Civil Informática
UTFSM, Valparaíso, Chile
Usuario #294897 counter.li.org

_______________________________________________
Condor-users mailing list
To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/condor-users

The archives can be found at either
https://lists.cs.wisc.edu/archive/condor-users/
http://www.opencondor.org/spaces/viewmailarchive.action?key=CONDOR