The symptom of the CODA problems was that End Run failed every time. A
few seconds of data would come in, then no data. We couldn't wake the
ROCs up by ping or by switching event monitoring. The two times I
looked at the FASTBUS crates, it was the left ROC that was sleeping; the
SFI light pattern was the same as for our usual sleep (Red, Yellow and
Green2 lit).
When we looked at the failed End runs with xcefdmp, there were no
physics events. All the events had a funny tag (0x21?) instead of "type
1 physics event"; the contents of an event looked like FASTBUS ADC and
TDC data from one crate.
Getting out of all CODA-related processes and starting fresh didn't
help. Rebooting the FASTBUS crates didn't help. Rebooting dblast07
didn't help. Power-cycling both FASTBUS crates fixed the problem.
****************************************************************************
If End Run fails, you can try the usual Reset, Download, Prestart, Go.
But if it keeps failing, power-cycle the crates. Be sure to analyze the
runs soon after you take them (or at least look at them with xcefdmp) to
be sure you don't have unknown event types.
****************************************************************************
Karen
This archive was generated by hypermail 2.1.2 : Mon Feb 24 2014 - 14:07:28 EST