Overview | Releases | Download | Docs | Links | Help | RecentChanges

MergedBocScanErrors

The fist sign of trouble:

Histogram polling timeout

  grep -a -A10 -B10 "olling timeout" /usera/sctrod/logs/SCT/sctrod/SctApiServer.out 

  ^@[MDSP: rodConfiguration.c,   735]::
  Configuring ROD: calibration mode.
  Histogram polling started
  Polling histogramming task
  scan point count 40
  Next bin 1+162 on ROD (0, 0, 0)
  Next bin 2+358 on ROD (0, 0, 0)
  Histogram polling timeout
  Expecting: 39 400
   ROD (0, 0, 0):  Event 358    Bin 2
  Event errors reported at time: Fri Nov 19 18:10:29 2004
   to 0xbf3ff848
  pollHistogramming complete (-1)
  Histogramming complete, 12 seconds, tidying up...
  Reading histograms...
  Modules in group 0 count: 1
  Module:  2 20220170100028
  SEND DATA from dsp 0 says ptr 0xa00e2000 length 0x52000
  --
  Next event 33446 (bin 0) on ROD (0, 0, 0)
  Next event 39961 (bin 0) on ROD (0, 0, 0)
  Next event 46475 (bin 0) on ROD (0, 0, 0)
  Next event 53118 (bin 0) on ROD (0, 0, 0)
  Next event 59631 (bin 0) on ROD (0, 0, 0)
  Next event 610 (bin 0) on ROD (0, 0, 0)
  Next event 7124 (bin 0) on ROD (0, 0, 0)
  Next event 13639 (bin 0) on ROD (0, 0, 0)
  Next event 20156 (bin 0) on ROD (0, 0, 0)
  Next event 24464 (bin 0) on ROD (0, 0, 0)
  Histogram polling timeout
  Expecting: 20 1000
   ROD (0, 0, 0):  Event 24464    Bin 0
  Event errors reported at time: Fri Nov 19 18:21:57 2004
  to 0xbf3ff848
  pollHistogramming complete (-1)
  Histogramming complete, 18 seconds, tidying up...
  Reading histograms...
  Modules in group 0 count: 1
  Module:  2 20220170100028
  Got some text: crate: 0 rod slot: 5

--

Source in SctApi/SctApiHisto?.cxx

  ScanControlRODHisto::reportTimeout()

produces "Scan aborted (histogramming stalled)" message to MRS.

---

Thanks to Bruce's suggestion, have figured out that above probs caused by error in implementation of RawScan causing doScan rather than doRawScan to be called. Now get new problem:

http://www.hep.phy.cam.ac.uk/~atlasdaq/images/screen.bug2.png

---

OK .. found other errors and have commited them to cvs. (22/11/2004) Now I can run a RawScan, and even my new RxThresholdBasedOnConfigRegister? and I get a rather suprising picture:

http://www.hep.phy.cam.ac.uk/~atlasdaq/images/strangeRxTrhesh.png

which it subsequently turned out was the kind of picture you get if you have too many print statements between successive horiziontal passes in the boc scan. When the print statement, and more importantly the delays caused by them, are removed. The picture finally looked normal.