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

Re: [HTCondor-users] submission to Condor CE from a remote submit Condor node



Hi Thomas,

This is actually the expected behavior for grid universe jobs: the grid job is submitted into your client schedd, which in turns spawn a gridmanager daemon that submits a vanilla job into the remote schedd (in this case the CE schedd). So these are just different representations of the same job and in the case of a CE in front of an HTCondor pool, you'll have a third instance of the job in your local HTCondor pool:

https://htcondor-ce.readthedocs.io/en/latest/overview/#on-htcondor-batch-systems

- Brian

On 7/16/20 8:52 AM, Thomas Hartmann wrote:
Hi again,

is it actually (easily?) possible to submit from a remote submit node to
a Condor CE?

bkg: I am trying to submit jobs based on [1] to a HTCondorCE. I.e.,
instead of a 'local' vanilla universe job, I want to submit the job as a
grid-type job to a remote CE defined by 'grid_resource'. However, the
jobs end up on the 'local' sched/cluster.

I guess the main problem is, that the submit node is actually a remote
submitter with the actual schedd being on another host [3].

Since we have several job transforms on the actual schedds, I suppose
one would need a transform there to divert grid-type jobs and forward
them to the Condor CE, or?

Cheers,
  Thomas


[1]
cat HTCondorCE.submit
universe = grid
use_x509userproxy = true
# +Owner = undefined
grid_resource = condor grid-vm08.desy.de grid-vm08.desy.de:9619
executable = mypayload.sh
output = stdout
error = stderr
log = logs
ShouldTransferFiles = YES
WhenToTransferOutput = ON_EXIT
queue


[2]
condor_submit -debug HTCondorCE.submit
07/16/20 15:44:43 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
07/16/20 15:44:43 Failed to decode JWT in keyfile
'/afs/desy.de/user/h/hartmath/.condor/tokens.d/x509up_u26551'; ignoring.
07/16/20 15:44:43 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
Submitting job(s)07/16/20 15:44:43 CREDMON: invoking
/var/lib/condor/util/get_token_cmd
07/16/20 15:44:44 CREDMON: storing cred for user hartmath@(null)
07/16/20 15:44:44 STORE_CRED: In mode 'add'
07/16/20 15:44:44 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
07/16/20 15:44:44 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
07/16/20 15:44:44 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
07/16/20 15:44:44 Can't open directory "/etc/condor/passwords.d" as
PRIV_UNKNOWN, errno: 2 (No such file or directory)
.
1 job(s) submitted to cluster 7188018.

[3]
grep HOST /etc/condor/config.d/10rsubmit.conf
CONDOR_HOST = bird-htc-master01.desy.de
SCHEDD_HOST = bird-htc-sched11.desy.de
CREDD_HOST = $(SCHEDD_HOST)

This body part will be downloaded on demand.