Wednesday bug triage report (2022-07-26)

Bryce Harrington bryce.harrington at
Fri Jul 29 23:17:29 UTC 2022

On Fri, Jul 29, 2022 at 10:23:15AM -0300, Athos Ribeiro wrote:
> - (New)            [dnsmasq]           - dnsmasq often using 100% of CPU
>   I tagged this as server-triage-discuss. Multiple users confirmed the bug, but
>   finding a reliable reproducer may be tricky. Instead of adding this to the
>   backlog, I am wondering if we have a clear path for this sort of (performance
>   issues) bug.

I googled on "dnsmasq is caught in a loop after every dns request,
resulting in 100% CPU".  In reading the suggested links (listed below),
here's some additional info that could be asked for, that might help
shed light:

  * dpkg --get-selections | cut -f 1
  * /etc/resolv.conf
  * /etc/NetworkManager/NetworkManager.conf
  * /etc/dnsmasq.conf
  * /run/dnsmasq/resolv.conf
  * /var/run/dhcpd.leases (or /run/dhcpd.leases?)

when the system is at 100%:
  * ps aux
    - Is logrotate running?  (Or some other 'heavy' process?)
  * journalctl -f
    - Any mention of "maximum number of concurrent dns queries reached"?

Possible tool to help reproduce

Unfortunately "Randomly 100% CPU" isn't ever really descriptive enough
to be actionable as a bug.  Could it just be a misconfiguration in the
dns settings?  This type of issue may be more effective to tackle first
as a tech support request through other channels, and if it can be
boiled down to a specific, reproducible problem (with a test case) then
handle as a proper bug.

The 'me-too' comment #3 could likely be having an unrelated problem and
should be reported as a different bug (with full logs and config)



More information about the ubuntu-server mailing list