 |
 |
This is an e-mail summary of production status by Kaushik De. Wei Yang | [log in to unmask] | 650-926-3338(O) Kaushik De wrote: > Hi Bob, > As Yuri and Tadashi reported, we are still having DQ2 > problems. I guess migration to a stable production version is > still ongoing. You can see this from the Panda error summary > for the past 12 hours: > > Job wall time: 23359 hrs Error losses: trans: 59 (0.3%) > panda: 106 (0.5%) ddm: 9284 (39.7%) other: 154 (0.7%) > > First, job wall time is half of what we see normally (jobs are > only really running at BNL, since that does not require DQ2 > transfers - files are all local). Second, 40% of jobs failed > because of DQ2 errors. > > Another indicator is the current status summary, which shows > 1159 assigned jobs at Tier 2's, but only 2 activated, and only > 370 running. We are not getting input files transferred to > Tier 2 sites. > > Finally, eveyone already knows about the issue with > transferring jobs. There are 14,595 in this state. Which > means, we cannot get files back from Tier 2 to BNL. > > So, we are hamstrung with DQ2 issues across the board. > I don't have any estimate of when we will be able to resume > production at the same level we used be able to with DQ2 0.2! > We continue to be in crises mode, since the migration. > Cheers, > Kaushik