When sending a SIGHUP to the unbound process it trys to reload its config and fails: Jan 21 12:44:05 bart unbound: [32746:0] info: service stopped (unbound 1.4.21). Jan 21 12:44:05 bart unbound: [32746:0] info: server stats for thread 0: 0 queries, 0 answers from cache, 0 recursions, 0 prefetch Jan 21 12:44:05 bart unbound: [32746:0] info: server stats for thread 0: requestlist max 0 avg 0 exceeded 0 jostled 0 Jan 21 12:44:05 bart unbound: [32746:0] notice: Restart of unbound 1.4.21. Jan 21 12:44:05 bart unbound: [32746:0] fatal error: Could not read config file: /unbound.conf .. because it is missing the path to its config file. For example the "openresolv" package makes use of this feature with the result that DNS will not work anymore until unbound is restarted manually everytime the DNS config changes (by dhcpd or openconnect for example). (Because resolvconf is called which in turn will try to reload unbound via SIGHUP after writing a new config for it) Current workaround is to specify a wrong PID file for unbound in /etc/resolvconf.conf so that resolvconf will restart the service instead of sending a SIGHUP to it: # mschiff: workaround until unbound will not die anymore # trying to find /unbound.conf as config when it is reloaded by SIGHUP # (resolvconf will restart the service if pid file is not found...) unbound_pid="/does/not/exist"
I recent version of unbound this must be resolved by not chrooting unbound using the 'chroot: ""' setting. Closing.
