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

Re: [Condor-users] weird reason for held job



I'm going to guess it is this...

https://condor-wiki.cs.wisc.edu/index.cgi/tktview?tn=1361

What's really happening is the job's Iwd doesn't exist on the execute machine.

Best,


matt

On 12/15/2010 04:22 PM, Dennis Box wrote:

No, I prepared it on a linux box, the worker nodes are linux as well.

I should have added that if I condor_submit the .cmd file from an nfs
mounted directory it runs, if I create a subdirectory in /tmp and submit
the .cmd file from there I get the 'no such file or directory' hold
reason. The full path to the executable is in the .cmd file.

Dennis


On 12/15/10 3:04 PM, Ian Cottam wrote:
Did it start
#!/bin/bash
?

Did you prepare it under Windows, such that Linux sees
#!/bin/bash\r
?

-Ian



On 15/12/2010 20:38, "Dennis Box"<dbox@xxxxxxxx> wrote:

I can create a condor job which gets held almost immediately after
submission:
[dbox@gpsn01 ~]$ condor_q dbox


-- Submitter: gpsn01.fnal.gov :<131.225.67.70:60205> : gpsn01.fnal.gov
ID OWNER SUBMITTED RUN_TIME ST PRI SIZE CMD
2070.0 dbox 12/15 13:58 0+00:00:00 H 0 0.0
hello_world.sh_201

Looking deeper into why it is held:

[dbox@gpsn01 ~]$ condor_q -l 2070.0 | grep HoldReason
LastHoldReason = "Error from slot2@xxxxxxxxxxxxxxxx: Failed to execute
'/grid/fermiapp/lbne/condor-exec/dbox/hello_world.sh_20101215_135745_1_wra

p.sh'
with arguments 360: No such file or directory"
LastHoldReasonCode = 6
LastHoldReasonSubCode = 2


Here's the weird part: I ssh to the machine where the error occurs and
look at it, the file seems to be fine!
[dbox@gpwn002 ~]$ ls -la
/grid/fermiapp/lbne/condor-exec/dbox/hello_world.sh_20101215_135745_1_wrap

.sh
-rwxr-xr-x 1 dbox gpcf 800 Dec 15 13:57
/grid/fermiapp/lbne/condor-exec/dbox/hello_world.sh_20101215_135745_1_wrap

.sh


Any suggestions on how to proceed to debug this?

Thanks
Dennis