Bug 1400874

Summary: Failed to add rhvh to engine via dhcp or static bond configured by NM during anaconda
Product: [oVirt] vdsm Reporter: dguo
Component: GeneralAssignee: Edward Haas <edwardh>
Status: CLOSED CURRENTRELEASE QA Contact: dguo
Severity: high Docs Contact:
Priority: medium    
Version: ---CC: bugs, cshao, danken, dguo, edwardh, gklein, huzhao, jcoscia, jiawu, jwoods, leiwang, mburman, weiwang, yaniwang, ycui, yzhao
Target Milestone: ovirt-4.1.0-betaKeywords: TestBlocker
Target Release: 4.19.1Flags: rule-engine: ovirt-4.1+
rule-engine: blocker+
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: 2017-03-16 14:50:44 UTC Type: Bug
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: Network RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Bug Depends On:    
Bug Blocks: 1304509, 1326798    
Attachments:
Description Flags
host_deploy log
none
network-scrpits before adding to engine
none
network-scripts after adding to engine
none
/var/logs under host
none
network-scripts and deploy logs
none
Include engine.log and vdsm log none

Description dguo 2016-12-02 08:37:48 UTC
Created attachment 1227214 [details]
host_deploy log

Description of problem:
Add rhvh to engine over dhcp bond failed, and the bond were configured during the anaconda installation.

Version-Release number of selected component (if applicable):
redhat-virtualization-host-4.0-20161130.0.x86_64
imgbased-0.8.10-0.1.el7ev.noarch
vdsm-4.18.17-1.el7ev.x86_64
Red Hat Virtualization Manager Version: 4.0.6-0.1.el7

How reproducible:
100%

Steps to Reproduce:
1. Install rhvh via anaconda
2. On the network page, add a dhcp bond contains two slave (em1 and p4p2) 
3. After the installation, add rhvh to engine over this bond

Actual results:
1.After step#3, Failed to configure management network on the host


Expected results:
1. After step#3, Add to engine successfully

Additional info:

Comment 1 dguo 2016-12-02 08:41:28 UTC
Created attachment 1227217 [details]
network-scrpits before adding to engine

Comment 2 dguo 2016-12-02 08:43:25 UTC
Created attachment 1227218 [details]
network-scripts after adding to engine

Comment 3 dguo 2016-12-02 08:44:10 UTC
Created attachment 1227219 [details]
/var/logs under host

Comment 4 Ying Cui 2016-12-07 07:40:01 UTC
Daijie, could you recheck this bug whether it is still here in last RHVH 4.0.6 build, as we knew, the previous build redhat-virtualization-host-4.0-20161130.0 built incorrect vdsm version in, see bug 1401205.

Comment 5 dguo 2016-12-07 09:32:46 UTC
Still caught the issue in the build redhat-virtualization-host-4.0-20161206.0, will attach the new logs

Comment 6 dguo 2016-12-07 09:33:48 UTC
Created attachment 1228949 [details]
network-scripts and deploy logs

Comment 8 dguo 2016-12-08 07:26:38 UTC
Besides, host can not be added to engine with the static VLAN configured during anaconda installation. Guess it is a same issue.

Comment 9 Edward Haas 2016-12-08 09:19:04 UTC
Please provide vdsm and supervdsm logs.
Perhaps even Engine logs will help here.

Comment 10 dguo 2016-12-08 10:00:27 UTC
Created attachment 1229428 [details]
Include engine.log and vdsm log

Please see the detail log

Comment 11 Dan Kenigsberg 2016-12-08 11:09:12 UTC
Which bond mode are you using? this could be a dup of bug 1367261

Comment 12 dguo 2016-12-09 08:36:03 UTC
Yes, the same mode (active-backup) with that bug 1367261. And before adding to engine, the bond's ip is did different with the ip got during anaconda.

Comment 13 Ying Cui 2016-12-11 15:26:58 UTC
(In reply to dguo from comment #8)
> Besides, host can not be added to engine with the static VLAN configured
> during anaconda installation. Guess it is a same issue.

Daijie, could you provide the detail steps if you failed adding host with static VLAN configured by NM after anaconda installation?

Comment 14 dguo 2016-12-12 02:23:58 UTC
I re-test with another two systems, adding to engine can be successful with static vlan confiugred during anacnda installation, so just ignore the Comment 8.

Comment 15 dguo 2016-12-12 11:12:29 UTC
Re-test the static bond configured by NM during anaconda installation is also failed to add to engine. 

I will share my test env by email

Comment 16 Dan Kenigsberg 2016-12-12 13:25:45 UTC
The limitation found in Vdsm is as follows: we cannot add a host if it has a bond, and a bond's slave has two connections: an active connection as a slave, and another inactive connection as an independent NIC.

Comment 24 Edward Haas 2016-12-20 21:13:07 UTC
Acquiring NetworkManager ifaces while NetworkManager is operational should solve this issue.

Please note that the host deploy currently defaults to disable NM service, therefore this patch will not solve this bug scenario.
It should be tested while setting the host deploy to leave NM running.

Comment 27 dguo 2017-02-27 06:30:18 UTC
Verified on rhvh-4.1-20170116.0(vdsm-4.19.1-1.el7ev)