nplan autopkgtests are failing in artful (NM 1.8?)
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
nplan (Ubuntu) |
Fix Released
|
High
|
Mathieu Trudel-Lapierre | ||
Xenial |
Fix Released
|
High
|
Mathieu Trudel-Lapierre | ||
Zesty |
Fix Released
|
Undecided
|
Unassigned | ||
Artful |
Fix Released
|
Undecided
|
Unassigned |
Bug Description
[Impact]
New versions of NetworkManager require more precise handling of the network interfaces. If not already configured to be ignored, bridges may be managed if no previous configuration exists, even if NetworkManager has not created the bridge.
[Test case]
Run autopkgtests for netplan:
autopkgtest -U --apt-pocket=
[Regression Potential]
Given that this only affects autopkgtests, regressions would be limited to the tests themselves. If the tests regularly fail to pass, or fail more than previously, this would be a regression caused by the update.
---
nplan's autopkgtests are failing in artful which is preventing network-manager 1.8 from migrating in to artful.
Maybe something changed in the new version of NetworkManager?
Tests are passing for armhf and s390x but failing for amd64, i386 and ppc64el.
Changed in nplan (Ubuntu): | |
status: | New → In Progress |
importance: | Undecided → High |
assignee: | nobody → Mathieu Trudel-Lapierre (cyphermox) |
Changed in nplan (Ubuntu): | |
status: | In Progress → Fix Committed |
description: | updated |
Changed in nplan (Ubuntu Xenial): | |
status: | New → In Progress |
importance: | Undecided → High |
assignee: | nobody → Mathieu Trudel-Lapierre (cyphermox) |
nplan was clearly not cleaning up after itself in the tests, and left behind a br0 device that NM would try to manage (and just spin forever, because there is no IP to be had there, and no DHCP server, etc).
There could still be a real regression in NM though; seems like it really should not be managing bridges/bonds it didn't create. We filed a bug upstream about that some time ago, maybe that changed back again?