The first injection made by Robert knocked the interferometer out of lock.  The second involved turning on a 'white noise' source.  The actual turning on of this source sent a glitch through the system.  The following is the time line of the glitch in the channels in which is has been observed:
 

                                Channel                                                    GPS time                                Time Delay

PEM-LVEA_SEISZ, SEI-LVEA_SEIS_Z                    732255534.2                                     0

PEM-LVEA_SEISX, SEI-LVEA_SEIS_X
PEM-BSC1_MIC, PEM-BSC3_MIC,                              732255534.7                                     0.5
PEM-HAM4_MIC

PEM-BSC1/2/3_ACCX, PEM-HAM2/3_ACCX         732255535.0                                      0.8
LSC-POB_I/Q onset

LSC-POB_I/Q maximum, LSC-AS_DC                        732255535.3                                     1.1

LSC-AS_Q, LSC-AS_I,                                                     732255535.4                                      1.2
DARM_CTRL, MICH_CTRL                            and         732255535.5                                      1.3
 

Thoughts for Duck
        - eventual successor to Donald and Ivana

Something on the following general lines general lines:

If (1) a seismic glitch is observed, and within 1 s a glitch is detected on an accelerometer,
           and within a  further 0.5 s glitches are observed on a mode cleaner channel and n₁ of the LSC channels,
           or within x s glitches are observed on n₂ LSC channels;

or (2) a glitch is observed on an accelerometer channel, followed with 0.y s by a suitable combination of
            mode cleaner and/or LSC channels;

or (3) glitches are observed within 0.z s on n₃ of the LSC channels;

then veto, say, 1 s from the time of onset of the first glitch on an LSC channel.

This would take care of the problem that any of the otherwise useful channels may at times begin glitching by itself for reasongs unknown - possible sometimes associated with alignment.  In the course of investigations in the reverse direction, begining from event candidates found by the inspiraltemplates, POB_I and POB_Q have been found to be useful as environmental indicators, but they are prone to glitching by themselves, without any other channel being affected.  This is also true of the REFL channels.

I haven't studied the WFS and optical lever channels , but they could possibly fit in (2):
            ... accelerometer or WFS or optical lever channel ...