jujud and mongo not restarted after a controller-config on one controller

[Benjamin Allot]

Hello,

trying to debug JAAS gce issues, I was told to apply the following setting:

juju controller-config mongo-memory-profile=low

on two different set of controllers

However, in one environment, one of the jujud and mongo were not restarted, resulting in them being isolated from the rest of the cluster upon reboot.

You can see it in the syslog attached. the setting is set at 15:09:52.
It had to wait until 15:26 until I manually restart the jujud-machine-0 service.

[Benjamin Allot (ballot)]

The log of another unit, where mongo and jujud restarted.

[Tim Penhey (thumper)]

Hi @ballot.

Just to be clear, some controllers did restart as expected? If that is the case, it is likely to be a bug in shutdown. We still have some shutdown races despite our best efforts.

If this happens grab a goroutine dump (juju_goroutines) from the controller that didn't shut down along with an egine report (juju_engine_report). This will help us identify where the blockage is.

[Benjamin Allot (ballot)]

Yes, this is exactly that. 2 controllers restarted, one did not.

Sadly I saw your message just after I restarted it on another environment.
I will do it if it happens again.

[Canonical Juju QA Bot (juju-qa-bot)]

This bug has not been updated in 2 years, so we're marking it Low importance. If you believe this is incorrect, please update the importance.