crash when trying to run on condor

[Sébastien Wertz]

Hi,

I'm trying to launch a very simple production on a condor cluster, with the latest MG5 version.

I changed the me5_configuration to have:

#! Default Running mode
#! 0: single machine/ 1: cluster / 2: multicore
run_mode = 1

#! Cluster Type [pbs|sge|condor|lsf|ge|slurm|htcaas|htcaas2] Use for cluster run only
#! And cluster queue (or partition for slurm)
cluster_type = condor
cluster_queue = madgraph

I got the following error at the end of the run:

Command "generate_events run_01" interrupted with error:
KeyError : 'cluster_tmp_path'

Included is the log file generated. I didn't see this problem when using v2.2.1, with the same options in the config file.

Cheers,
Sébastien

[Kenneth Long (kdlong-e)]

Hi Sébastien,

I encountered this error as well. The fix is easy: in <process dir>/bin/internal/madevent_interface.py", line 2513, in do_combine_events function:

 2513 if self.options['run_mode'] ==1 and self.options['cluster_tmp_path']:

change cluster_tmp_path to cluster_temp_path. It's just a typo for the correct variable name in the run card. I'm not sure if this has been fixed yet in a release, but I did point it out to the authors in a forum question a week or so ago.

Cheers,

Kenneth

[Sébastien Wertz (sebastien-wertz)]

Hi Kenneth,

Thanks, I didn't notice the typo!

But to get it to work I also had to modify madgraph/interface/madevent_interface.py, event *after* I had created the process directory (and modified the file you pointed at)...

Anyway, it works now, thanks again!