Today we saw the second instance of what seems like a serious guardian bug, where the guardian executes the line of code above a return True statement, but doesn't return True and exit the state. Screen shots of both incidents are attached. After the first inicident (shown in the first attached screenshot) Jamie suggested that possibly the requested state could have been erroneously set to be the state LOCK_DRMI_1F.
Today I am sure that the requested state was DRMI_ON_POP, so it should have changed states after returning true.
I'm investigating:
https://bugzilla.ligo-wa.caltech.edu/bugzilla3/show_bug.cgi?id=830
Sheila, please provide as much information as you can, in the bug report, on what exactly you tried to get out of the problem. Did you re-request anything? pause/un-pause? MANUAL? How did you eventually get out of the situation?
In general, when reporting bugs please provide as much information as possible. It's much easier to debug the situation when all relevant information is provided.
This could be my fault, I incremented the counter which I should not have done in the last step.