Hi Michael,
I'm certainly the least expert in this case - I was just trying to help mapping 
that other case I have seen.

Your ouput tells (me oO) that your resolvd talks to:
1. a link scope DNS 192.168.2.254 on wlp4s0
2. is set up to talk to a gloabl dns at 127.0.0.1

IIRC it will ask both and whoever answers first is the reply it gives
(can be bad at times if the nack of server A is faster than the detail
from server B).

On 127.0.0.53 that is resolvd itself listening.

The libvirt dnsmasq usually has "except-interface=lo" so it should not listen 
on 127.0.0.1.
The case I mentioned in c#25 was a manual config change that made resolved call 
the dnsmasq causing the loop.
I agree that you can't easily derive if you are "looped" now from the output 
you have.

Maybe you could check in your case if/who does listen on 127.0.0.1:53?
$ sudo netstat -ltaupn | sed -rne 2p -e '/:53\b/p'

Does libvirt know the pid of the dnsmasq it spawned, maybe this can be used to 
make this check more find grained?
Like "is any of our pids binding :53 on something in systemd-resolved --status" 
maybe.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1670959

Title:
  systemd-resolved using 100% CPU

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/dnsmasq/+bug/1670959/+subscriptions

-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to