state: support state:ignore

[cathay4t]

Introducing InterfaceState.IGNORE for:
* The ignored interface will not be changed when applying desire state.
* The state verification will not be impacted by ignored interface.

Test cases added.

[packit-as-a-service]

Congratulations! One of the builds has completed. 🍾

You can install the built RPMs by following these steps:

• sudo yum install -y dnf-plugins-core on RHEL 8
• sudo dnf install -y dnf-plugins-core on Fedora
• dnf copr enable packit/nmstate-nmstate-1298
• And now you can install the packages.

Please note that the RPMs should be used only in a testing environment.

[codecov]

Codecov Report

Merging #1298 into base will increase coverage by 0.18%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             base    #1298      +/-   ##
==========================================
+ Coverage   57.91%   58.10%   +0.18%     
==========================================
  Files          67       67              
  Lines        5316     5332      +16     
==========================================
+ Hits         3079     3098      +19     
+ Misses       2237     2234       -3     

Impacted Files	Coverage Δ
libnmstate/nm/profile.py	25.13% <ø> (ø)
libnmstate/ifaces/base_iface.py	96.84% <100.00%> (+0.02%)	⬆️
libnmstate/ifaces/ifaces.py	90.90% <100.00%> (+1.74%)	⬆️
libnmstate/schema.py	100.00% <100.00%> (ø)

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 83fd948...146b6e3. Read the comment docs.

[ffmancera]

I am fine with the current approach but I feel like it could be confusing to use state to mark the interface as ignore. Because ignore it is not a state itself. Maybe it would be better to add a new parameter that could be mark:ignore or something like that, indeed that could be useful to mark the interfaces in the future, what do you think?

[EdDev]

I am fine with the current approach but I feel like it could be confusing to use state to mark the interface as ignore. Because ignore it is not a state itself. Maybe it would be better to add a new parameter that could be mark:ignore or something like that, indeed that could be useful to mark the interfaces in the future, what do you think?

Adding another field is odd. I think that when applying, "ignore" means... "do not do UP, DOWN or DISABLE" which translates to not touching it. So it seems ok to me.

When reporting, it will never be IGNORE, right? This merits a comment in the commit, on the API and in documentation.

[cathay4t]

I am fine with the current approach but I feel like it could be confusing to use state to mark the interface as ignore. Because ignore it is not a state itself. Maybe it would be better to add a new parameter that could be mark:ignore or something like that, indeed that could be useful to mark the interfaces in the future, what do you think?

Adding another field is odd. I think that when applying, "ignore" means... "do not do UP, DOWN or DISABLE" which translates to not touching it. So it seems ok to me.

When reporting, it will never be IGNORE, right? This merits a comment in the commit, on the API and in documentation.

The ignore is only used when applying desire state.
The use case is RHV/oVirt is expecting nmstate/NM don't touch the TUN/TAP interface managed by libvirt.

Currently, nmstate don't touch them as they are unknown to us. But I think the correct way is mark them as ignored in desire state.

[ffmancera]

As everyone agrees on introducing state: ignore I will not block on this. Approved, thanks!

[cathay4t]

QE approval request sent. Please wait test feedback before merge it.

[smyle10]

Applied 'state: ignore' to main port and subordinate port respectively, it works well.

[cathay4t]

Timeout on waiting comsumer QE. Merging.