Hello again all,
It seems I'm good for nothing today and I mucked up a few links in my previous mail, for Sheffield and QM (at least they're the ones I noticed).
The corrected links are below, and on the bulletin.
Sorry about that all.
Matt
SHEFFIELD
https://ggus.eu/index.php?mode=ticket_info&ticket_id=114460 (18/6)
Gridpp (and other) VO pilot roles at Sheffield. No news for a while, snoplus are trying to use pilot roles now for dirac so this is becoming very relevant. In progress (9/10)
https://ggus.eu/index.php?mode=ticket_info&ticket_id=116560 (30/9)
Sno+ jobs failing, likely due to too many being submitted to the 10 slots that Sno+ has. Maybe a WMS scheduling problem - Stephen B has given advice. Elena asked if the problem persisted a few weeks ago. Waiting for reply (12/10)
https://ggus.eu/index.php?mode=ticket_info&ticket_id=116967 (17/10)
A ROD availability ticket, on hold as per SOP. On hold (20/10)
QMUL
https://ggus.eu/index.php?mode=ticket_info&ticket_id=117151 (23/10)
A ticket about jumbo frame problems, submitted to QM. After Dan provided some education the user replied, in that he only sees this problem at two atlas sites. But he is contacting the network admins at his institution to see if it is their end. On hold (29/10)
https://ggus.eu/index.php?mode=ticket_info&ticket_id=117011 (19/10)
ROD ticket for glue-validate errors. Went away for a while after Dan re-yaimed his site bdii, but possibly back again. Daniela suggests re-running the glue-validate test. Reopened (2/11)
https://ggus.eu/index.php?mode=ticket_info&ticket_id=116689 (6/10)
Another ROD ticket, where Ops glexec test jobs are seemingly timing out for QM (this is the ticket Daniela mentioned on the ops mailing list). Dan noted that with the cluster half full[1] tests were passing, suggesting some kind of load correlation (but as he also notes - what's getting loaded and causing the problem - Batch, CE or WNs?). Kashif reckons the argus server, and suggests a handy glexec time test which he posted. In progress (2/11)
|