You’re right. I’m trying to figure out the best way to handle this, since it does seem that the timer is never re-armed after that line. The timer is initially scheduled in amt_rcv_relay_mem_query, then disarmed on that line. I’ll have to brush up on data racing.