Changes between Version 37 and Version 38 of PS1_IPP_Czarlog_20120716

Timestamp:

Jul 20, 2012, 1:06:41 AM (14 years ago)

Author:

Mark Huber

Comment:

--

PS1_IPP_Czarlog_20120716

@@ -85 +85 @@
  * 21:10 ippc17 high load, restarting pstamp and update
  * 19:30 stdscience and stack restarted once SAS stacks and LAP diffs finished. Tried boosting stack processing but chip/warp processing irregular so restarting stdscience+stack.
- * 22:00 Chris suspects the warp stage/stsci load trouble is due to the ecliptic not having @HOST@ in the workdir.  LAP runs do have this defined, with 300 warps running only see some load spikes, ~50, but clear quickly. 
- * xx:xx watch out for when the ecliptic chips finish and all nodes go to warp in ecliptic pantasks (~23:00)
+ * 22:00 Chris suspects the warp stage/stsci load trouble is due to the ecliptic not having @HOST@ in the workdir.  LAP runs do have this defined, with 300 warps running only see a few load spikes, ~50, but clear quickly. 
+ * 22:30 watch out for when the ecliptic chips finish and all nodes go to warp in ecliptic pantasks (~23:00) -- stopped ecliptic pantasks for now, ~2k warps left, to not conflict with main pantasks (be careful of the node balance when in use, commented out multiple compute3).
+ * 23:30 LAP stacks back on after chip+warps seen can reach 100-150. +1x compute3 to stack since deepstack not using tonight.
 === Friday : 2012.07.20 ===