Hi Steve, It looks like you're using the whole time that the job is in the "Running" state from the logging info. This includes the sandbox stage-in time (but bizarrely not the stage-out time), so it is probably just a slow network link into the WN/VM network. (For more accurate user payload timing it's best to only measure the "Job_<number> | Running" states). There isn't much more we can investigate about this as we don't get any pilot logs from VAC/vcycle sites (I don't know whether there is any plan to make these available in future versions, if such a log even exists on these nodes). Regards, Simon On Mon, Mar 14, 2016 at 05:12:53PM +0000, Steve Lloyd wrote:
Hi, If I submit jobs with no destination to Dirac sometimes they go to CLOUD.CERN-PROD.ch which seems to send them to vcycle at Manchester. The interesting thing is these are 'Hello World' and typically take ~12secs executing. The ones going to CLOUD.CERN-PROD.ch take more like 1m 20sec - many times longer. Presumably this is some startup overhead that one wouldn't notice in a real job? See http://pprc.qmul.ac.uk/~lloyd/gridpp/gridtests/uktest.html for a summary. Cheers +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ Prof Steve Lloyd Head of School of Physics and Astronomy Queen Mary University of London, Mile End Road, London E1 4NS, UK E-mail: s.l.lloyd@qmul.ac.uk Phone: +44-(0)207-882-6967 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++