Reinstalling nvidia drivers requires a reboot

Bug #1991010 reported by Chris Johnston
6
This bug affects 1 person
Affects Status Importance Assigned to Milestone
Containerd Subordinate Charm
Fix Released
Medium
Unassigned

Bug Description

When the nvidia drivers are reinstalled either through the upgrade-packages action or through the charm, the worker will need to be rebooted in order to use the GPU. We should surface this need to the users by setting a message (and possibly blocked status) in juju status.

This requirement also highlights the issue found in LP#1982894 where changing the sources or the packages will always cause a reinstall. There should be a check added to see if there is a change in what should be installed prior to uninstalling/reinstalling.

Additionally, we probably shouldn't be autoremoving stuff that isn't relevant to what we are trying to do. For example, I have seen a kernel get autoremoved through this functionality. We should attempt to limit the autoremoval to just what our needs are.

Example:
https://pastebin.ubuntu.com/p/NwcrQGkk4h/

description: updated
Adam Dyess (addyess)
Changed in charm-containerd:
status: New → Confirmed
importance: Undecided → Medium
milestone: none → 1.26
Revision history for this message
Adam Dyess (addyess) wrote :

Provides a warning on upgrade that a reboot may be needed:
https://github.com/charmed-kubernetes/charm-containerd/pull/80

George Kraft (cynerva)
Changed in charm-containerd:
status: Confirmed → Fix Committed
Adam Dyess (addyess)
Changed in charm-containerd:
status: Fix Committed → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.