Bug details: ** https://bugzilla.redhat.com/show_bug.cgi?id=2185849 ** Information from BlockerBugs App:
The votes have been last counted at 2023-04-11 18:25 UTC and the last processed comment was #comment-851126
To learn how to vote, see: https://pagure.io/fedora-qa/blocker-review A quick example: BetaBlocker +1 (where the tracker name is one of BetaBlocker/FinalBlocker/BetaFE/FinalFE/0Day/PreviousRelease and the vote is one of +1/0/-1)
BetaBlocker +1
BetaBlocker
FinalBlocker
BetaFE
FinalFE
0Day
PreviousRelease
+1
0
-1
(removed vote updated comment below)
It's a little confusing, but I don't see it being a violation. It is the RC, which is bit-for-bit what we ship. If it labeled as a beta, that would definitely be a problem. I'm weak on the -1, though, because to someone who doesn't understand how our process works, this could make them think they don't have the GA artifacts.
FinalFE +1
Worth fixing if we have to re-spin for something else.
My worry was a bit different. I'm not sure if we're going to rename the files for GA. Let's consider both.
If we rename the files: Internally, the image has some partition label used for installation. That shouldn't change if we just rename the file, and so the installation shouldn't be affected. But, compose metadata might refer to those RC files and if we rename them, tools processing those metadata might not be able to find the right images. (I have no idea what those tools can be). Also, in general, I feel a bit uneasy testing something which is not yet in a final form, but declaring it gold.
If we don't rename the files: Some people might be confused, because we don't usually put RC in image names (and just some image names). But more importantly, again, some tools might fail to find the images if they're suddenly named differently. For example, will Fedora Media Writer find them without issues? This is quite hard to test, because QA can make local patches to test it, but in real infra the config files might get generated completely differently. What about virt-install, Boxes, etc. There might be some issues caused by the inconsistent RC suffix.
I wonder if somebody familiar with all these bits can say "this is not a problem", we just take the risk, or we avoid the risk by re-spinning it :-)
I'm still a little reluctant to vote for the blocker status by now
FinalFE +1 FinalBlocker -1
This definitely wouldn't pass the "last blocker at Go/No-Go" smell test for me.
I think we'd have to rename the files, otherwise we'd need to make a lot of edits to the website in order for people to be able to download them, which seems more error prone than a one-liner to rename the files.
The question is: can we rename the files? I'll let @kevin and @humaton weigh in on that. If they say "we can't" or "we can, but it's very unpleasant" then I'll be +1 to block on it just from the logistics standpoint.
I'm a stong blocker here.
We cannot easily 'rename' things without breaking the entire compose in many ways. All the checksum files will be wrong. All the metadata will be wrong. Looking for things in koji will not match up. Internal things in the images may well refer to the name with rc.
If we don't rename them, forever people will ask 'what does RC mean?'
FinalBlocker +1
Who am I to argue with Kevin?
@kevin is right :-)
FinalBlocker +1 AGREED AcceptedFinalBlocker
The following votes have been closed:
Metadata Update from @blockerbot: - Issue status updated to: Closed (was: Open)
Release F38 is no longer tracked by BlockerBugs, closing this ticket.
Log in to comment on this ticket.