Marking for fast track for @adamwill's proposal, and also +1 to marking it an F40 final blocker.

Just for the records, here is the ticket https://pagure.io/fedora-qa/blocker-review/issue/1593

Hmm, so the bug is already marked as FinalBlocker through the normal voting procedure. I don't think FESCo needs to get involved. I hope the maintainers reply in the bug and we have a clearer situation, but that might not happen before Monday.

No, it's only proposed as such. Technically there's not a good basis for it unless we designate it as such.

OK, in that case I can +1 to marking this is a blocker so that we don't release before making a decision. But it'd still be good to hear from the maintainers before committing to a particular course of action.

yeah, I would not count the blocker vote as valid, tbh. I specifically said in the proposal that I was proposing it for consideration by FESCo and it does not violate any criteria. Technically it didn't need to be proposed as a blocker for that, but I didn't want us to forget about it.

So, network has been on life support for quite a while, but I agree it's retirement should have more notice.
I don't think just that openqa is using it is enough, but I think thats a pretty good indicator that there are people using it and they need time to migrate. :)

So, +1 to making this a fesco blocker.

FESCo folks, it would be awesome if you could decide this today or Sunday rather than waiting for the meeting Monday, so that if you do decide to make this a blocker we can get a new initscripts built and a compose done. If we wait till the meeting the compose won't be done till Monday night west coast time, which is putting us back in a bit of a testing crunch. thanks!

Hmmm. I found this surprising too. However, it also sparked some deja vu, so I went digging.

The F33 change "NetworkManager keyfile instead of ifcfg rh" says:

By now, initscripts' network-script package is deprecated in Fedora and upstream wants to move away from that format in the long term.

That seems to not have made the release notes for F33, but there was a follow-up change for F36, which did get into those release notes — although it doesn't actually spell out "deprecated". There's also a corresponding Fedora Magazine article: https://fedoramagazine.org/converting-networkmanager-from-ifcfg-to-keyfiles/

And then... there's this from 2018 https://lists.fedoraproject.org/archives/list/users@lists.fedoraproject.org/thread/753G6H4BAF23WTG4IOA2X5YFKRN7NBSK/, where someone noticed that the script itself prints a deprecation warning.

And then, there's this! https://fedoraproject.org/wiki/Changes/MigrateIfcfgToKeyfile -- an F39 change which says

With this change, existing profiles in ifcfg format will be automatically migrated to the native keyfile format via a migration service shipped with the NetworkManager-initscripts-ifcfg-rh package. In Fedora, we plan to drop the plugin by Fedora Linux 41.

That doesn't seem to have gotten any devel-list conversation (no replies at all), but there was a little bit of chatter in the FESCo ticket.

"by F41" does technically include "F40"....

Side note: I think this could perhaps be a 0-day update (bringing network-service back). It's not shipped on any media. Likely anyone still using it would install via network or get it from updates.

"by F41" does technically include "F40"....

I'm not sure. I'm pretty sure that many people will interpret "by F41" as "by the F41 changes complete deadline" or even later. This is very ambiguous phrasing for something that seems quite important.

@jistone, @dcantrell, @humaton, @sgallagh, @mhayden, @tstellar: your vote please

More importantly, that is not referring to this.

"the plugin" referred to in that quote is the NetworkManager plugin that reads configuration files in ifcfg format. It is not /etc/init.d/network, which is what this ticket is about. They are two different things.

I already acknowledged in the original description that the network service "was in some ways marked as deprecated", but I don't think that's sufficient to justify removing it without further notice or warning.

edit: I also agree with @kevin that this doesn't need to make the media, so we won't hold the RC for it. 0-day should be fine.

This was discussed during the FESCo meeting on 2024-04-15:
AGREED: proposal: FESCo asks that the legacy network subpackage be reinstated for F40. The bug requesting the reinstatement is marked as F40-0day and F40-freeze-exception. APPROVED(+7,0,0) AGREED: Proposal: the network-scripts should block F40 GA. REJECTED(+1,0,-6)

The package was reinstated in time for F40.