Hi!
Most items in the report are self-explanatory (it's the same as the weeks before ...), so I will just add a comment on the ARDA Dashboard Summary on slide number 6.
In the grid section you find a large number of terminated jobs with status 'aborted' or 'unknown', accumulated during the two weeks.
This leads to a grid efficiency of about 70% only for the job robot and 60% for the production jobs, which is quite low.
I did not investigate this further, so I do not know what the reasons are. Does anyone else know? Admins? Seems like the situation improved again during the last days.
A large fraction of applications had an unknown exit code, what has already been observed by Christian the week before.
Most of the failures for job robot applications are due to a bug in the application, see
https://hypernews.cern.ch/HyperNews/CMS/get/sc4/594.html
benedikt
There are minutes attached to this event.
Show them.
