I have a heterogeneous pool, including IA64 and X86_64. The X86_64 server is the submitter, and others are work nodes. I compiled my source file on worker nodes, and submited it from the submitter. Submiting my job, I used condor_q to query job, and the result is as follows.
ID OWNER SUBMITTED RUN_TIME ST PRI SIZE CMD
7046.0 zhxue 9/14 17:41 0+00:27:19 R 0 0.0 ia64
7047.0 zhxue 9/14 17:41 0+00:00:01 H 0 0.0 data
Why 7047.0 is generated?
Furthermore, I use "condor_q -analyze" command, and it prompts the following:
7046.000: Request is being serviced
7047.000: Request is held.
Hold reason: Error from starter on slot2@**.**.**: Failed to execute '/home/zhxue/.globus/.gass_cache/local/md5/58/1da5713002eb7a2d6fe3f76e3f673a/md5/b5/f7f0ea2e16e03c4fdb16fcbbb5abd9/data': Exec format error
It seems 7047.0 is the execution process, but it can not been scheduled to IA64 servers. (slot2@**.**.** is a core with x86_64 architecture).
I specified "requirements" in the submit script, but it seems not work. The script is as follows:
grid_resource = gt2 ***.***.***:/jobmanager-condor
requirements = Arch == "IA64" && OpSys == "Linux"
output = ......
log = ......
Would you like to help me? Any suggestion is appreciated. Thanks.