From Renee, 03.12.12:
https://twist.phys.ualberta.ca/forum/view.php?bn=twist_software&key=1071269872

Posted by Renee Poutissou (renee@triumf.ca) , Friday Dec 12, 2003 15:38

Bad runs in this context means: " there seems to be a lost of
synchronization between the event numbers in the two independant
fastbus crates".

A test for this condition exist in MOFIA giving the messages
"Suspicious pulser ...." This test can have false positive due to some
other problem and the data is still good.

When I studied runs with the MOFIA messages last year, I correlated
the real synchronisation errors to a message given by the Frontend
PPCs " X. Error... " This error ws just written to the log file
midas.log.  Since the start of running this fall, this error stops the
run immediately.

The attached file contains all the Bad runs since the beginning of
time. The safest thing is to ignore these runs completely. If we loose
too much data, it is possible to look at the event number where this
happened and back off by 1000 events. All events prior should be OK.

INSTRUCTION FOR FUTURE USE

To build this list, there is a script in /data_onl/current
check_Xerrors.csh.  The midas.log file is renamed regularly to keep
the file to a manageable size.

Ex.  To get the X. errors in file
        midas.log.2003_08_26_to_12_
you have to be user twistonl

>cd /twist/data_onl/current

>,/check_Xerrors.csh midas.log.2003_08_26_to_12_12 

Output will be in file Xerr_midas.log.2003_08_26_to_12_12