Bugzilla – Bug 1135150
dnsmasq fails to start on boot
Last modified: 2019-07-15 13:28:04 UTC
Since resolv.conf is a link to be recreated by wicked, dnsmasq fails to start on boot pretty reproducibly: ● dnsmasq.service - DNS caching server. Loaded: loaded (/usr/lib/systemd/system/dnsmasq.service; enabled; vendor preset: disabled) Drop-In: /run/systemd/generator/dnsmasq.service.d └─50-insserv.conf-$named.conf Active: failed (Result: exit-code) since Wed 2019-05-15 07:17:37 CEST; 2min 22s ago Main PID: 3590 (code=exited, status=5) May 15 07:17:37 cleopatra systemd[1]: Starting DNS caching server.... May 15 07:17:37 cleopatra dnsmasq[3570]: dnsmasq: syntax check OK. May 15 07:17:37 cleopatra dnsmasq[3590]: dnsmasq: directory /etc/resolv.conf for resolv-file is missing, cannot poll May 15 07:17:37 cleopatra dnsmasq[3590]: directory /etc/resolv.conf for resolv-file is missing, cannot poll May 15 07:17:37 cleopatra dnsmasq[3590]: FAILED to start up May 15 07:17:37 cleopatra systemd[1]: dnsmasq.service: Main process exited, code=exited, status=5/NOTINSTALLED May 15 07:17:37 cleopatra systemd[1]: dnsmasq.service: Failed with result 'exit-code'. May 15 07:17:37 cleopatra systemd[1]: Failed to start DNS caching server.. After boot everything is fine, so a simple restart fixes it.
Fails also on 15.1.
Unfortunately the error message is misleading. See bug 1123699 comment 34 for a detailed explanation of what's going on. *** This bug has been marked as a duplicate of bug 1123699 ***
No, this has nothing to do with libvirt.
The other one was found in a libvirt context, but the underlaying problem that dnsmasq fails to start because the directory that contains /etc/resolv.conf does not exist is the same. *** This bug has been marked as a duplicate of bug 1123699 ***
Andreas, please test if this helps: # echo "d /run/netconfig 0755 root root -" > /etc/tmpfiles.d/netconfig.conf # reboot I suggested this in bug 112369 comment 36, but nobody seems to have tried it yet.
Yes, this works for me.
Thanks for testing. I just noticed that I missed a digit in the bug number I posted in comment 5. I meant bug 1123699 comment 36.