Reports until 16:25, Tuesday 01 September 2026
H1 CDS
david.barker@LIGO.ORG - posted 16:25, Tuesday 01 September 2026 (91779)
CDS Maintenance Summary: Tuesday 1st September 2026

WP13560 CAL CS remove channels

Joe, Erik, Jonathan, Dave:

We installed a new h1calcs model which removes channels from the DAQ. We verfified none of these channels were being sent by the DAQ broadcaster.

The rev-update showed a surprising number of modified mdl files used in the build, details are:

Changed dependency revisions in new version:
cal/common/models/CAL_CS_MASTER.mdl : r28532 --> r35947
cal/common/models/CAL_LINE_MONITOR_MASTER.mdl : r28567 --> r35947
cal/h1/models/h1calcs.mdl : r27506 --> r35949
sus/common/models/ESD_LINEARIZATION_WITH_CHARGE_MASTER.mdl : r16336 --> r30316

Removed dependencies in new version:
cds/common/src/ATAN2.c
 

The INI file just changed by the removal of the channels:

--: slow channel H1:CAL-CS_DARM_ARM_LENGTH_AVG removed from the DAQ 
--: slow channel H1:CAL-CS_DARM_ARM_LENGTH_X removed from the DAQ 
--: slow channel H1:CAL-CS_DARM_ARM_LENGTH_Y removed from the DAQ 
--: slow channel H1:CAL-STRAIN_SLOW removed from the DAQ 
--: fast channel H1:CAL-STRAIN_DBL_DQ removed from the DAQ 
 

sw-lvea-aux crash, reboot, build replacement

Richard, Fil, Erik, Jonathan, Dave:

As noted in other alogs, a lot of today's time was taken investigating and recovering the CER main POE switch which drives the majority of the gige-cameras. Erik is building a spare Cisco Catalyst 3560 switch as a swap-out replacement.

WP13569 Guardian reboot

TJ, Ryan S, Erik:

h1guardian1 was rebooted for PM and to clear its CFC flags. No problems with this reboot.

WP13578 Add Guardian nodes to DAQ

TJ, Dave:

H1EPICS_GRD.ini was modified to add the two new guardian nodes CRS_HAM3 and SPIH23_STAT (+46 chans). EDC+DAQ restart was required.

DAQ Restart

Erik, Jonathan, Dave:

The DAQ was restarted for the new h1calcs model and EDC additional grd channels.

This did not go well.

( A Reminder: in the new DAQ, the framewriters continue to run throughout, switching to the new configuration automatically. Only the DC, NDS and TW restart)

At 11:13 I restarted the 0-leg. This caused both FW0 and FW2 to crash. We then restarted the EDC, but the framewriters did not come back.

Erik restarted the FW daqd services, noting these now have new names (rts-daqd_rs.service)

Jonathan looked at the 0-leg logs, we determined there was nothing we could do to prevent FW1 from failing, we would just recover it afterwards.

At 11:41 we restarted the 1-leg, and FW1 did not restart. It behaved normally, picking up the new run number and configuration.

Jonathan is investigating what happened to FW0/2 and why FW1 did not crash.