Bug 1596062

Summary:	Fedora-Cloud-Base - qcow2 - eth0 is not UP
Product:	[Fedora] Fedora	Reporter:	Bruno Goncalves <bgoncalv>
Component:	distribution	Assignee:	Josh Boyer <jwboyer>
Status:	CLOSED RAWHIDE	QA Contact:	Fedora Extras Quality Assurance <extras-qa>
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	rawhide	CC:	dan, deekej, dustymabe, john.j5live, jonathan, kdudka, kevin, lnykryn, normand, zbyszek
Target Milestone:	---	Keywords:	Regression
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2018-07-20 13:08:08 UTC	Type:	Bug
Regression:	---	Mount Type:	---
Documentation:	---	CRM:
Verified Versions:		Category:	---
oVirt Team:	---	RHEL 7.3 requirements from Atomic Host:
Cloudforms Team:	---	Target Upstream Version:
Embargoed:

Description Bruno Goncalves 2018-06-28 08:01:22 UTC

Description of problem:
After updating initscripts to initscripts-9.83-1.fc29.x86_64 the qcow2 image does not bring eth0 up.

Version-Release number of selected component (if applicable):
initscripts-9.83-1.fc29.x86_64

How reproducible:
100%

Steps to Reproduce:
1. Download qcow2 image:
https://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-Rawhide-20180615.n.0/compose/Cloud/x86_64/images/Fedora-Cloud-Base-Rawhide-20180615.n.0.x86_64.qcow2
2. bring it up and update initscripts there.
dnf install --disablerepo=* https://kojipkgs.fedoraproject.org//packages/initscripts/9.83/1.fc29/x86_64/initscripts-9.83-1.fc29.x86_64.rpm
3. reboot
4. during boot you can see eth0 is not UP:

         Starting Initial cloud-init job (metadata service crawler)...
[   82.819590] cloud-init[493]: Cloud-init v. 17.1 running 'init' at Thu, 28 Jun 2018 07:50:19 +0000. Up 81.83 seconds.
[   82.830228] cloud-init[493]: ci-info: +++++++++++++++++++++++++++Net device info++++++++++++++++++++++++++++
[   82.840308] cloud-init[493]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+
[   82.850239] cloud-init[493]: ci-info: | Device |   Up  |  Address  |    Mask   | Scope |     Hw-Address    |
[   82.859962] cloud-init[493]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+
[   82.869918] cloud-init[493]: ci-info: | eth0:  | False |     .     |     .     |   .   | 52:54:00:12:34:56 |
[   82.879968] cloud-init[493]: ci-info: |  lo:   |  True | 127.0.0.1 | 255.0.0.0 |   .   |         .         |
[   82.890433] cloud-init[493]: ci-info: |  lo:   |  True |     .     |     .     |   d   |         .         |
[  OK  ] Started Initial cloud-init job (metadata service crawler).
[   82.900678] cloud-init[493]: ci-info: +--------+-------+-----------+-----------+-------+-------------------+



Actual results:
 eth0:  | False |

Expected results:
| eth0:  | True | 10.0.2.15 | 255.255.255.0 |   .   | 52:54:00:12:34:56 |

Additional info:
it works well with initscripts-9.82-1.fc29.x86_64

Comment 1 Lukáš Nykrýn 2018-06-28 10:19:41 UTC

The network-scripts were split from the initscripts package. In your case, you just specified that you want to only upgrade the main package, so ended up without network-scripts, but the general "dnf upgrade" path works. So I don't think that this is a problem:

https://paste.fedoraproject.org/paste/DtR7Xnr9VQtrhhHtIseK3Q

But, the fact that cloud images still use network-scripts is the actual bug here. There was a fesco decision, that they should be moved to networkd years ago.

These days I would be more for NetworkManager, but definitely not the old scripts. This needs to be investigated and fixed.

Comment 2 Dusty Mabe 2018-06-28 12:49:13 UTC

I'm thinking we should just add in NetworkManager in the cloud base image since that's what the rest of Fedora uses.

Comment 3 Bruno Goncalves 2018-06-28 13:05:35 UTC

Note, I've reported this issue due https://pagure.io/cloud-sig/issue/289

Comment 4 Kevin Fenzi 2018-06-28 15:51:58 UTC

(In reply to Dusty Mabe from comment #2)
> I'm thinking we should just add in NetworkManager in the cloud base image
> since that's what the rest of Fedora uses.

Yeah, I think NetworkManager would be fine too, although some might complain that it leaves a process running taking up ram.

Comment 5 Bruno Goncalves 2018-07-12 08:06:53 UTC

I'm just wondering if there is any update about this issue as Fedora-Cloud-Base-Rawhide-20180711.n.0 still does not work.

Comment 6 Dusty Mabe 2018-07-12 13:20:20 UTC

I have asked jdoss to open a PR to pagure.io/fedora-kickstarts on master branch to add NM to the cloud base image.

please ping back in a week.

Comment 7 Bruno Goncalves 2018-07-16 10:42:09 UTC

Thanks for the PRs, I do not know if it is related to it or not, but the current rawhide qcow2 is not booting at all.

With Fedora-Cloud-Base-Rawhide-20180711.n.0.x86_64.qcow2 I see it booting up.

but on Fedora-Cloud-Base-Rawhide-20180714.n.0.x86_64.qcow2 I do not see any message of boot process.

Comment 8 Bruno Goncalves 2018-07-20 07:07:27 UTC

I've just tried Fedora-Cloud-Base-Rawhide-20180719.n.0.x86_64.qcow2 and it works for me. Do you think this BZ can be closed?

http://kojipkgs.fedoraproject.org/compose/rawhide/Fedora-Rawhide-20180719.n.0/compose/Cloud/x86_64/images/Fedora-Cloud-Base-Rawhide-20180719.n.0.x86_64.qcow2

Comment 9 Dusty Mabe 2018-07-20 13:08:08 UTC

sure! :)