Bug 1585462 - [HE Wizard] - Add proper error message when trying to deploy HE on top of bond modes=0,5,6
Summary: [HE Wizard] - Add proper error message when trying to deploy HE on top of bon...
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: ovirt-ansible-collection
Classification: oVirt
Component: hosted-engine-setup
Version: unspecified
Hardware: x86_64
OS: Linux
medium
medium
Target Milestone: ovirt-4.4.6
: 1.4.0
Assignee: Aviv Turgeman
QA Contact: Wei Wang
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2018-06-03 09:09 UTC by Michael Burman
Modified: 2021-05-14 07:30 UTC (History)
10 users (show)

Fixed In Version: ovirt-ansible-collection-1.4.0
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2021-05-14 07:30:23 UTC
oVirt Team: Integration
Embargoed:
aturgema: needinfo-
rule-engine: ovirt-4.4+
weiwang: testing_plan_complete+
mtessun: planning_ack+
sbonazzo: devel_ack+
yzhao: testing_ack+


Attachments (Terms of Use)
record (989.37 KB, application/x-gzip)
2018-06-03 09:09 UTC, Michael Burman
no flags Details
mode6_error_cockpit (6.66 KB, image/png)
2018-08-03 06:47 UTC, Yihui Zhao
no flags Details
naming conventions for VLAN and bonds (156.82 KB, image/png)
2021-01-04 12:21 UTC, Aviv Turgeman
no flags Details
bond mode 0 (154.68 KB, image/png)
2021-01-04 12:22 UTC, Aviv Turgeman
no flags Details
existing networks available (173.99 KB, image/png)
2021-01-04 12:23 UTC, Aviv Turgeman
no flags Details
network naming info label (188.85 KB, image/png)
2021-02-11 09:18 UTC, Aviv Turgeman
no flags Details
network naming info label correct position in UI (186.21 KB, image/png)
2021-02-11 09:22 UTC, Aviv Turgeman
no flags Details


Links
System ID Private Priority Status Summary Last Updated
Github oVirt ovirt-ansible-collection pull 226 0 None open role: hosted_engine_setup: Filter VLAN devices with bond parent 2021-03-01 13:41:46 UTC

Description Michael Burman 2018-06-03 09:09:44 UTC
Created attachment 1447146 [details]
record

Description of problem:
[HE Wizard] - Add proper error message when trying to deploy HE on top of bond modes=0,5,6.

As part of the fix for BZ 1578418, the HE deploy shouldn't start if the host active connection is on top of bond modes=0,5,6, this modes are not supported for the ovirtmgmt bridge.

After this fix the HE won't start in cockpit and fail with too generic error:
'System data could not be retrieved!'

This error should be changed to something more useful to understand, like the bond mode is not supported for HE deploy, or like on CLI
'A Network interface is required'
The current error is too much generic

Version-Release number of selected component (if applicable):
0.11.24-1.el7ev.noarch
cockpit-165-3.el7.x86_64
ovirt-hosted-engine-setup-2.2.22-1.el7ev.noarch

How reproducible:
100%

Steps to Reproduce:
1. Try to run HE via cockpit on top of a bond=0,5,6

Actual results:
generic error:
'System data could not be retrieved!'

Expected results:
More useful error to the user

Additional info:
See also 1578418

Comment 1 Ryan Barry 2018-06-03 12:54:00 UTC
Does this work on the CLI?

Comment 2 Michael Burman 2018-06-03 13:02:11 UTC
(In reply to Ryan Barry from comment #1)
> Does this work on the CLI?

It fails on CLI as well as expected. In both ways, HE shouldn't start if the bond is mode=0,5,6 and has the host active connection. 

Cockpit - 
IF the host active connection is on top of a bond mode=0,5,6, the HE wizard won't be loaded and we failing with a very generic error:

'System data could not be retrieved!'

Can't start the HE deploy in such scenario, but we must to handle this generic error, will report a fresh bug to handle this. 

CLI-
If the host active connection is on top of a bond mode-0,5,6, the HE deploy is failed:
[ ERROR ] Failed to execute stage 'Environment customization': A Network interface is required
[ ERROR ] Hosted Engine deployment failed

Failure here is the expected behaviour, but the generic error on cockpit side should be improved.

Comment 3 Yihui Zhao 2018-08-03 06:47:14 UTC
Created attachment 1472885 [details]
mode6_error_cockpit

Comment 4 Yihui Zhao 2018-08-03 06:50:40 UTC
Can reproduce.

Use version:
rhvh-4.2.5.1-0.20180801.0+1
ovirt-hosted-engine-ha-2.2.16-1.el7ev.noarch
ovirt-hosted-engine-setup-2.2.25-1.el7ev.noarch
cockpit-ovirt-dashboard-0.11.31-1.el7ev.noarch
rhvm-appliance-4.2-20180801.0.el7.noarch


Test steps:
1. Configure the network with mode 6
2. Deploy HE via cockpit.

Result:
The same as the description.
mode6_error_cockpit

Comment 5 Aviv Turgeman 2021-01-04 12:21:41 UTC
Created attachment 1744308 [details]
naming conventions for VLAN and bonds

Can't reproduce

Version tested:
ovirt-node-ng-installer-4.4.4-2020121712.el8
cockpit-ovirt-dashboard-0.14.17-1.el8.noarch
ovirt-hosted-engine-setup-2.4.9-1.el8.noarch
ovirt-hosted-engine-ha-2.4.5-1.el8.noarch
ovirt-release44-pre-4.4.4-0.6.rc6.el8.noarch

Steps taken:
1. Configured bonded network interface with mode 0 balance-rr (Round Robin)
2. Started HE installation via cockpit, recieved following massage boxes:
A. "System data could not be retrieved!"
B. "No valid network interface has been found
If you are using Bonds or VLANs Use the following naming conventions:
- VLAN interfaces: physical_device.VLAN_ID (for example, eth0.23, eth1.128, enp3s0.50)
- Bond interfaces: bond*number* (for example, bond0, bond1)
- VLANs on bond interfaces: bond*number*.VLAN_ID (for example, bond0.50, bond1.128)
* Supported bond modes: active-backup, balance-xor, broadcast, 802.3ad
* Networking teaming is not supported and will cause errors"

screenshot of error is attached to this comment.

Comment 6 Aviv Turgeman 2021-01-04 12:22:32 UTC
Created attachment 1744309 [details]
bond mode 0

Comment 7 Aviv Turgeman 2021-01-04 12:23:09 UTC
Created attachment 1744310 [details]
existing networks available

Comment 8 Sandro Bonazzola 2021-01-08 07:28:58 UTC
Moving to QE as per comment #5 this should be already fixed.

Comment 9 Wei Wang 2021-01-12 05:55:22 UTC
Test Version:
rhvh-4.4.4.1-0.20210106.0
cockpit-ws-224.2-1.el8.x86_64
cockpit-224.2-1.el8.x86_64
cockpit-ovirt-dashboard-0.14.17-1.el8ev.noarch
cockpit-bridge-224.2-1.el8.x86_64
cockpit-system-224.2-1.el8.noarch
cockpit-storaged-224.2-1.el8.noarch
subscription-manager-cockpit-1.27.16-1.el8.noarch
cockpit-dashboard-224.2-1.el8.noarch

Test step:
1. Setup bond2(using mode 0 rr) with cockpit UI
2. Go to hosted engine deployment to start hosted engine deploying

Result:
Cannot get the right massage boxes in cockpit UI, the HE wizard displays.

Maybe I used wrong test steps. Could dev help to correct me? thanks

BTW:
I also test vlan over the bond2(using mode 0 rr) to deploy hosted engine, HE installation is failed, but still no the right massage boxes in cockpit UI.

Comment 10 RHEL Program Management 2021-01-14 07:00:15 UTC
Target release should be placed once a package build is known to fix a issue. Since this bug is not modified, the target version has been reset. Please use target milestone to plan a fix for a oVirt release.

Comment 11 Wei Wang 2021-01-14 07:00:44 UTC
According to comment 9, move it to "ASSIGNED"

Comment 12 Aviv Turgeman 2021-02-11 09:18:41 UTC
Created attachment 1756343 [details]
network naming info label

(In reply to Wei Wang from comment #9)

> Test step:
> 1. Setup bond2(using mode 0 rr) with cockpit UI
> 2. Go to hosted engine deployment to start hosted engine deploying
> 
> Result:
> Cannot get the right massage boxes in cockpit UI, the HE wizard displays.
Hi Wei,

I got 2 different situations when trying to reproduce:
1. My env is connected only to unsupported bond mode 'bond0', in this case I received the proper message boxes, and cockpit-UI wizard is not displaying at all.

2. When using more then one connections to my env, 'bond0' as unsupported bond mode and 'ens3' , in this case cockpit-UI wizard is displaying without showing box-messages and when trying to choose a network,
the bond is not appearing in the list of networks we can choose from, which is the expected behavior, but could be confusing.

My suggestion (2.):
Adding to cockpit-UI an info icon that can clearify naming conventions as constant, (I've added an example of how it might look on wizard)


> Maybe I used wrong test steps. Could dev help to correct me? thanks
> 
> BTW:
> I also test vlan over the bond2(using mode 0 rr) to deploy hosted engine, HE
> installation is failed, but still no the right massage boxes in cockpit UI.

that make sense since still using the unsupported bond, but IMO the vlan is also needed to be filtered out, since using the unsupported bond mode

Comment 13 Aviv Turgeman 2021-02-11 09:22:39 UTC
Created attachment 1756345 [details]
network naming info label correct position in UI

Comment 14 Sandro Bonazzola 2021-04-16 07:21:03 UTC
https://github.com/oVirt/ovirt-ansible-collection/commit/2f49d4edb944d37ead39eccad1529ca85c654782 is included in ovirt-ansible-collection-1.4.0

Comment 15 Wei Wang 2021-04-16 08:20:44 UTC
The latest RHVH build RHVH-4.4-20210413.0-RHVH-x86_64-dvd1.iso has https://bugzilla.redhat.com/show_bug.cgi?id=1946095 issue, so QE will verify this bug after BZ 1946095 is fixed.

Comment 17 Wei Wang 2021-05-12 02:14:55 UTC
Test with RHVH-4.4-20210510.0-RHVH-x86_64-dvd1.iso (ovirt-ansible-collection-1.4.2-1.el8ev.noarch)

Test scenarios
1. My env is connected only to unsupported bond mode 'bond0'
Yes, I got the right proper message boxes, and cockpit-ui wizard is not displaying at all. 

2. When using more then one connections to my env, 'bond0' as unsupported bond mode and 'ens3'.
Yes, I got the expected result.
Cockpit-UI wizard is displaying without showing box-messages and when trying to choose a network, the bond is not appearing in the list of networks we can choose from, but there is a prompt of help to show the message boxes.

Bug is fixed, move it to "VERIFIED"


Note You need to log in before you can comment on or make changes to this bug.