A number of webCoRE users have been reporting issues with timer based pistons for the last couple of weeks, not just since DST, though one user recently tentatively suggested that it seemed that the problems may suddenly have lifted. Often the pistons have not been running at all and the recovery procedures have kicked in. I’ve also seen one which had multiple semaphore waits for no apparent reason.
A semaphore wait means the piston has been triggered while it is already running and the new execution has backed off for ten seconds to give the original time to finish. It is perfectly normal (though perhaps not ideal) if multiple triggers can fire round about the same time , or if the piston changes something that will cause it to be triggered again. It is not normal if the piston is only being triggered by a single ‘happens at’ timer.
The semaphore wait happened at 6:50:38 am. The piston seems to have run at 6:50:30 am. It would be interesting to know if it actually triggered late, or if it was completely missed and it was the recovery procedure that ran it.
webCoRE has a recovery procedure that looks out for pistons that failed to execute when scheduled and executes them. Without this the piston’s next run might not get scheduled.