At Andy's request, I have compiled a list of psinject start/stop times that occurred
during observing mode, based on the times of starts, stops and errors reported in
the psinject log files on h1hwinj1. For inserting segments in the database, I have
substracted 1 second from the start time and added 1 second to the end time.
I have also verified that I can see a large glitch in H1:CAL-INJ_HARDWARE_OUT_DQ
at these times:

1130155228 1130155231 Oct 29 2015 12:00:11 UTC
1130155254 1130155257 Oct 29 2015 12:00:37 UTC
1132130708 1132130711 Nov 21 2015 08:44:51 UTC
1132130741 1132130744 Nov 21 2015 08:45:24 UTC
1132162463 1132162466 Nov 21 2015 17:34:06 UTC
1132162498 1132162501 Nov 21 2015 17:34:41 UTC

The tcl script used to find these times is attached, along 
with a concatenation of the psinject log files read by the script.