| 14 | | === Thursday : YYYY.MM.DD === |
| 15 | | |
| | 14 | === Thursday : 2017.08.31 === |
| | 15 | * MEH: Weryk using nodes c23-c63, x065-x088, s00-s14 for tracklet linking again -- primarily CPU use, nice'd played nice with normal stack and other processing previously but keeping of of primary nightly science and data nodes |
| | 16 | * past 24 hours seems ok except for c23-c28 -- ppSub and stamp jobs stalled, possibly on file access -- as the c23-c25 (c0) and c26-c28 (part c2) are not part of normal processing (as-needed only), commenting them out in ~ippitc/ippconfig/pantasks_hosts.input |
| | 17 | * may be related to rsync stalling between compute nodes Weryk noticed yesterday when distributing 28GB file -- specific example ippc27 data to ippc42 (rate rapidly drops to few KB/s) |
