SIGUSR1 on a busy relay causes watchdog timeout
steps to reproduce:
- run a busy relay (5000+ connections I guess?)
- pkill -USR1 tor
expected results:
connection stats are printed relatively expeditiously or tor gives up
actual results:
19 connections printed in less than 1 second, then 50 seconds pass, then:
systemd[1]: tor.service: Watchdog timeout (limit 1min)!
systemd[1]: tor.service: Killing process 234 (tor) with signal SIGABRT.