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

Re: [Condor-users] How to specify no of machines for a Globus/condor-G job



On Mon, 23 Feb 2009, Samir Khanal wrote:

Hi Steve

I have version 7.0.5  of condor installed.
Its a ROCKS cluster 5.1


[skhanal@comet ~]$ cat ex1.sub
executable = /bin/hostname
Transfer_Executable = false
globusscheduler = protos.cs.bgsu.edu/jobmanager-pbs
#grid_resource = gt4 protos.cs.bgsu.edu PBS
universe = globus
output = date.out
error=date.error
log = date.log
queue


this does submit job and complete successfully.
I want this to run on multiple nodes.

Samir

The only way I know of to get it to run on multiple nodes
is to add a globusrsl line to specify
(jobtype=multiple)(xcount=3)
where 3 nodes is what you would want to run on.


Condor-C nowadays can bypass the GT4 interface and talk
directly to the pbs master too but I have no experience with that.

Steve Timm




________________________________________
From: condor-users-bounces@xxxxxxxxxxx [condor-users-bounces@xxxxxxxxxxx] On Behalf Of Steven Timm [timm@xxxxxxxx]
Sent: Monday, February 23, 2009 11:13 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] How to specify no of machines for a Globus/condor-G job

Yes, the grid univers manual should apply to that version of globus,
but you should update to the latest condor 7.2 since it has
special support for globus 4.2.  There is some incompatibiliyt
between condor 6.8 (which only supported up to globus 4.0) and globus 4.2.

That may not fix the problem below, however, but it is the
first thing to try.

Steve

On Mon, 23 Feb 2009, Samir Khanal wrote:

Hi Steve

FYI I have globus4.2.1

does this section http://www.cs.wisc.edu/condor/manual/v6.8/5_3Grid_Universe.html#SECTION00632400000000000000
apply to this version of globus?


[skhanal@comet ~]$ condor_q
-- Submitter: comet.cs.bgsu.edu : <129.1.64.81:37521> : comet.cs.bgsu.edu
ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD
119.0   skhanal         2/23 22:49   0+00:00:00 H  0   0.0  hostname
120.0   skhanal         2/23 22:49   0+00:06:19 R  0   0.0  gridftp_wrapper.sh

2 jobs; 0 idle, 1 running, 1 held
[skhanal@comet ~]$ condor_q -analyze
-- Submitter: comet.cs.bgsu.edu : <129.1.64.81:37521> : comet.cs.bgsu.edu
ID      OWNER            SUBMITTED     RUN_TIME ST PRI SIZE CMD
---
119.000:  Request is held.

Hold reason: Failed to create proxy delegation


my scrpit looks like

executable = /bin/hostname
Transfer_Executable = false
grid_resource = gt4 protos.cs.bgsu.edu PBS
universe = grid
output = date.out
error=date.error
log = date.log
queue


any pointers?
Samir
________________________________________
From: condor-users-bounces@xxxxxxxxxxx [condor-users-bounces@xxxxxxxxxxx] On Behalf Of Steven Timm [timm@xxxxxxxx]
Sent: Monday, February 23, 2009 9:53 PM
To: Condor-Users Mail List
Subject: Re: [Condor-users] How to specify no of machines for a Globus/condor-G job

First--change universe=globus to universe = grid
and globusscheduler=....
to
GridResource= gt2 hostname/jobmanager-pbs
-----------------------

then put in
globusrsl = (jobtype=multiple)(xcount=3)

The details of which classad attributes get sent
to Grid Universe jobs and which don't are in the Grid Universe
section of the condor manual.

Note--the implementations of jobmanager-pbs vary some from
grid to grid but the syntax I gave should be good
for OSG or Teragrid for sure.

Steve Timm


On Mon, 23 Feb 2009, Samir Khanal wrote:

Hi List

I have a condor job submission file
as following


executable = test
arguments=
Transfer_Executable = true
globusscheduler = protos.xx.xxx.xxx/jobmanager-pbs
machine_count = 3
universe = globus
output = test.out
error=test.error
log = date.log
queue



WARNING: the line `machine_count = 3' was unused by condor_submit. Is it a typo?

This works fairly well and prints 1 hostname.

I want this to run on multiple nodes,
How do i specify machine count in a globus environment?

Thanks
Samir
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/


--
------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/


--
------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/
_______________________________________________
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:
https://lists.cs.wisc.edu/archive/condor-users/


--
------------------------------------------------------------------
Steven C. Timm, Ph.D  (630) 840-8525
timm@xxxxxxxx  http://home.fnal.gov/~timm/
Fermilab Computing Division, Scientific Computing Facilities,
Grid Facilities Department, FermiGrid Services Group, Assistant Group Leader.