Bug 1786547

Summary: Cannot reapply when bridge profile previously does not contain 802-3-ethernet section
Product: Red Hat Enterprise Linux 8 Reporter: Gris Ge <fge>
Component: NetworkManagerAssignee: NetworkManager Development Team <nm-team>
Status: CLOSED WONTFIX QA Contact: Desktop QE <desktop-qa-list>
Severity: high Docs Contact:
Priority: high    
Version: 8.2CC: amusil, atragler, bgalvani, danken, dholler, lrintel, mburman, pasik, phoracek, rkhan, sukulkar, thaller, till
Target Milestone: rc   
Target Release: 8.0   
Hardware: x86_64   
OS: Linux   
Whiteboard:
Fixed In Version: Doc Type: If docs needed, set a value
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2020-08-24 06:08:46 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:
Bug Depends On:    
Bug Blocks: 1738136, 1785558    
Attachments:
Description Flags
Reproduce script none

Description Gris Ge 2019-12-26 07:32:56 UTC
Created attachment 1647662 [details]
Reproduce script

Description of problem:

When a bridge profiles does not have 802-3-ethernet section,
the reapply will failed after created 802-3-ethernet section with error:

    Can't reapply any changes to '802-3-ethernet' setting

even the newly created 802-3-ethernet section is identical to runtime status.

Version-Release number of selected component (if applicable):
NetworkManager-1.22.0-2.el8.x86_64

How reproducible:
100%

Steps to Reproduce:
1. ./nm_bug.sh


Actual results:

Error: Reapplying connection to device 'br0' (/org/freedesktop/NetworkManager/Devices/8) failed: Can't reapply changes to '802-3-ethernet.speed' setting

Expected results:

Reapply works.

Additional info:

Comment 2 Thomas Haller 2020-02-14 16:48:39 UTC
> Could you provide update on this bug and possible fix schedule

A fix is not planned for 8.2. So it will be 8.3 at earliest (unless planning changes).



Reapply is intended for small changes while avoiding disruption. If Reapply fails because the changes are too large, then the solution is always to do a full activation.

Now, you might say, "but in this example the change is not large". Yes, you are probably right. But it's is really just a missing feature where you cannot reapply certain small changes without disruption. The workaround is either don't do it or live with the disruption. That means, such a bug to Reapply doesn't seem seems high priority to me (in general).

You are welcome to provide arguments why it's high priority for nmstate, and we may agree to adjust planning. Please more specific than "caused a lot [of test] failure[s]". Thank you!

Comment 3 Gris Ge 2020-02-20 14:39:08 UTC
Hi Thomas,

Thanks for the update.

I will try to find a possible workaround in nmstate on this.
If failed, I will ask RHV and CNV team to state their impacted use case.

Comment 4 Gris Ge 2020-02-27 08:30:22 UTC
Will try to workaround NM issue in nmstate.

Tracking in upstream is: https://github.com/nmstate/nmstate/issues/868

Comment 5 Till Maas 2020-02-27 08:34:09 UTC
(In reply to Thomas Haller from comment #2)

> 
> You are welcome to provide arguments why it's high priority for nmstate, and
> we may agree to adjust planning. Please more specific than "caused a lot [of
> test] failure[s]". Thank you!

The RHV tests indicate that it is important for them not to have this disruption for their customer workloads. Dominik, could you please specify how important non-disruptive changes are for RHV?

Comment 6 Dominik Holler 2020-02-27 09:29:03 UTC
(In reply to Till Maas from comment #5)
> (In reply to Thomas Haller from comment #2)
> 
> > 
> > You are welcome to provide arguments why it's high priority for nmstate, and
> > we may agree to adjust planning. Please more specific than "caused a lot [of
> > test] failure[s]". Thank you!
> 
> The RHV tests indicate that it is important for them not to have this
> disruption for their customer workloads. Dominik, could you please specify
> how important non-disruptive changes are for RHV?

Very important, because this is the behavior like it was before RHV used NetworkManager and what the users are expecting.

Comment 7 Gris Ge 2020-03-25 15:35:20 UTC
Hi Dominik,

The nmstate-0.2.6-4.8.el8 in RHEL 8.2 has applied a workaround.

Actions like adding slaves to bridges/bond will not cause link disruption any more.

Please let me know if anything unexpected link disruption happens.

Comment 8 Ales Musil 2020-03-26 09:28:44 UTC
Ok,

it seems that the link is stable again. I will enable the stability monitor in our tests.

Comment 9 Petr Horáček 2020-08-22 13:05:46 UTC
There was no action on this bug in the last 5 months. Are there any plans to continue with the effort or are ok with the workaround and could close this?

Comment 10 Gris Ge 2020-08-24 06:08:46 UTC
The workaround used in nmstate looks good to me.

Closing.