LXD no longer activates all interfaces on initial deploy when using MAAS2rc3 and JUJU Beta13
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
Canonical Juju |
Fix Released
|
High
|
Unassigned |
Bug Description
When using MAAS 2RC3 and Juju 2 Beta 13 deploying charms to LXD on MAAS controlled machines only ETH0 activates. The ENI is perfect and if I issue sudo lxc restart juju-machine all of the interfaces are then activated. This is new to Beta 13.
Fresh deploy of a charm
juju deploy charm --to lxd:0
then from machine 0
ubuntu@
+------
| NAME | STATE | IPV4 | IPV6 | TYPE | SNAPSHOTS |
+------
| juju-970ebd-0-lxd-0 | RUNNING | 10.1.0.2 (eth0) | | PERSISTENT | 0 |
+------
ubuntu@
ubuntu@
+------
| NAME | STATE | IPV4 | IPV6 | TYPE | SNAPSHOTS |
+------
| juju-970ebd-0-lxd-0 | RUNNING | 10.1.0.2 (eth0) | | PERSISTENT | 0 |
| | | 10.10.0.2 (eth1) | | | |
| | | 10.20.0.2 (eth2) | | | |
| | | 10.2.0.2 (eth3) | | | |
+------
no changes made anywhere, just a restart container and it works. Makes deploying bundles and large scripts difficult since the networks are not activated and subordinate services fail to deploy until you manually restart the containers.
3 physical machines to recreate issue
1 - MAAS
2 - JUJU controller
3 - Physical server for services
Changed in juju-core: | |
status: | New → Triaged |
importance: | Undecided → Critical |
milestone: | none → 2.0-beta14 |
Changed in juju-core: | |
importance: | Critical → High |
no longer affects: | maas |
Changed in juju-core: | |
milestone: | 2.0-beta14 → none |
Changed in juju-core: | |
status: | Incomplete → Triaged |
milestone: | none → 2.0.0 |
Changed in juju-core: | |
milestone: | 2.0.0 → 2.0-beta15 |
affects: | juju-core → juju |
Changed in juju: | |
milestone: | 2.0-beta15 → none |
milestone: | none → 2.0-beta15 |
I have the same issue on MAAS 2.0rc3 and Juju 2.0-beta13 when deploying the openstack-base bundle.
The work around of restarting the containers fixes the issue, but is a hassle.