OVS metrics attempted to collect, but no sudoers and no ovs on the system

[Xav Paice]

Using charm release cs:telegraf-38, a unit with no OVS installed was reporting errors:

[inputs.exec] Error in plugin: exec: exit status 1 for command '/var/lib/juju/agents/unit-telegraf-prometheus-0/.venv/bin/python /var/lib/juju/agents/unit-telegraf-prometheus-0/charm/files/telegraf_exec_metrics.py --metric ovs_dpctl': sudo: no tty present and no askpass program specified...

The sudoers entry telegraf_sudoers was missing entirely on this unit. Even adding it, copied from another unit, wasn't enough to clear the error.

This particular unit then attempted to send an email for every exec that failed, and filled the disk.

[Alvaro Uria (aluria)]

Released at cs:~llama-charmers-next/telegraf-4

[Junien F (axino)]

Is it possible to backport the fix for this critical bug to the "regular" telegraf charm before the 20.10 release ?

Thanks

[Nikolay Vinogradov (nikolay.vinogradov)]

Seeing this on the deployment.

[Haw Loeung (hloeung)]

Same here, causing cron spam and bringing down our MX server.

[David Coronel (davecore)]

subscribed ~field-medium for awareness

[Haw Loeung (hloeung)]

Can we get the charm built with this fix and published?

[Andrea Ieri (aieri)]

This seems to be available in the 20.10 release:

$ git --no-pager branch -r --contains 2def7
  origin/HEAD -> origin/master
  origin/bug/1896467
  origin/master
  origin/stable/20.10