Ciao Alessandro,
> Thanks again Maarten, everything seems to be working now! The job output and
> the interaction with the RB is OK.
I submitted a few jobs as "ops" (local user ops001) and they got stuck;
I saw they ran in LSF, but the grid_monitor kept seeing them as pending.
In the end I cancelled them all, killed the remaining ops001 processes,
and moved ~ops001/.lcgjm out of the way: after that a new job went OK.
I also submitted a job as "opssgm" and it worked fine, except that the
environment on atlas-wn-204 is bad: the job gets a lot of YAIM variables
and configuration functions etc. in its environment! Please fix that.
|