Similar to the HEPI's done last week, I took the text file generated by Barker, which is extracted from the Guardian logs, and removed the truncated channels from that list. The MASTERSWITCH was also added to the list. This I used for the do not monitor channel list. I saved this file with the safe.snap files in /opt/rtcds/userapps/release/isi/h1/burtfiles.
Next the safe.snap is modified to include the monitor bit on all channels not RO and not fields. This is then loaded into the FE with the LOAD TABLE ONLY button.
After confirming a successful load and noting the DIFF list, the safe.snap is then modified with the donotinclude list. When theTABLE is LOADED again, most of the diffs go away and the remaining differences represent the differences we care about wrt most recent safe.snap file. Either the channel is set wrong or the snap needs updating.
All HAM-ISI are good except HAM3, there remains some blend diffs as we test configs for the 0.6hz peak issue.
I'll work on the BSC-ISI soon.
This exercise exposed a problem: if you have more than 40 diffs, it only shows 40 and it only lists 40. It makes you think there is only 40 when there are more. I've filed bug 796.
I have confirmed that the guardian reports (displays a notification) when the filter module is not in the correct state. This may not be true for other subsystem Guardians. Betsy and I saw this was the case on some ETM filter module that Dave's list reports that guardian touched. The monitoring that Guardian does for the SEI channels may need to be coded explicitly.
This has nothing to do with the new SDF functionality, nor has guardian behavior changed. The SEI guardian systems were programmed from the beginning to notify on changes to any of the controller filter banks. This is specific to SEI and is not being done by any of the other subsystems.