If I launch an instance from Fedora-Cloud-Base-25-1.3.x86_64 image, this instance is ssh-connectible and in its openstack log from dashboard I can see these lines regarding network configuration:
[ 5.510999] cloud-init[631]: ci-info: | eth0: | True | . | . | . | fa:16:3e:a3:cb:13 | [ 5.511152] cloud-init[631]: ci-info: | eth0: | True | . | . | d | fa:16:3e:a3:cb:13 |
If I, however, create a snapshot out of that instance and launch it, the resulting instance is not ssh-connectible (from copr-be-dev network) and I can see this line the in the openstack log:
[ 23.011290] cloud-init[612]: ci-info: | eth0: | False | . | . | . | fa:16:3e:93:3e:d3 |
Snapshot instances for f24 and f25ppc64le images do not seem to have this problem (tried on F24 Cloud Base release image and Fedora-Cloud-Base-25-1.2.ppc64le.qcow2 image respectively).
Can you retry your tests?
The controller node (fed-cloud09) had it's disk full, and also was (mis)configured to allow compute instances on it (it shouldn't have any normally). So it's possible one or both of those things caused this.
:checkered_flag:
The same error appeared. This is the log from the snapshot instance:
In coprdev tenant, there are currently two instances. One created from Fedora Cloud Base 25 (ip: 172.25.148.223) and another one from snaphost of that instance (ip: 172.25.148.224). The snapshost does cannot be connected to and again according to console log, eth0 is down. The whole log is attached.
I needed to reproduce the problem again with "f25cloud-base-instance" - 172.25.148.236 and "f25snapshot-instance" - 172.25.148.241 as the previous instances got deleted by vm cleaning script.
With f26 images, network in their snapshots works again.
Talking on irc it seems we can close this out now.
:postal_horn:
Metadata Update from @kevin: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)