Description of problem: If I modify the disk in some way in anaconda's blivet partitioning and then want to reset the changes, pressing "reset all" button doesn't do anything. In custom partitioning, "reset all" works fine. On F31 DVD, "reset all" worked fine even in blivet partitioning. Version-Release number of selected component (if applicable): Fedora-Server-dvd-x86_64-32-1.4.iso How reproducible: always Steps to Reproduce: 1. prepare an existing installation (I used a default Server minimal installation, using LVM+xfs) 2. boot Anaconda, go to Blivet partitioning 3. delete all standard partitions, delete all LVs, delete all LVMs -> have the disk completely clean 4. press "reset all" 5. nothing happens Actual results: Disk state doesn't reset Expected results: Disk state resets to actual on-disk state
Created attachment 1680522 [details] bug demonstration video See this video to see the bug occur.
Created attachment 1680523 [details] anaconda.log
Created attachment 1680524 [details] journal
Created attachment 1680525 [details] lvm.log
Created attachment 1680526 [details] program.log
Created attachment 1680527 [details] storage.log
Created attachment 1680528 [details] syslog
I'm not sure if this is a blocker or "just" a commonbug, but it seems severe enough, proposing for discussion.
It seems I can go back to the original state using "Undo last action" repeatedly. However, I'm not certain this will work for flawlessly for a high number of complex operations. Also, it only works as expected before you hit "Reset all". Even though it seemingly does nothing, I have found cases when subsequent "Undo last action" will then switch you to a different state than the one you arrived from. To sum up, as long as "Reset all" is broken, I'd consider any use of it to switch the system into an undefined state.
Created attachment 1680535 [details] video - reset all, then undo, not original state Here's an example of how "Reset all" and then "Undo last action" doesn't go back to the original state - the first partition used to be sda1, but then it is sda2. So it is not suitable as a workaround.
Fixed in a pull request: https://github.com/rhinstaller/anaconda/pull/2458
Created attachment 1680539 [details] An updates image with the fix
+1 blocker I think this is a visible enough problem early in the experience that it's worth blocking on, although I won't object if the consensus is that it's not a blocker because there's a workaround that can be documented in CommonBugs.
(In reply to Vendula Poncova from comment #12) > Created attachment 1680539 [details] > An updates image with the fix Fixes the problem, thanks.
Seems like a +1 Blocker to me.
@Kamil: Was this broken in F31 as well? If so, I'll offer this fix a +1 FE, but if it worked then and has regressed in F32, I'll agree to call it a blocker.
From comment0 > On F31 DVD, "reset all" worked fine even in blivet partitioning. It worked in F31, which is why I went with +1 Blocker.
+1 blocker
+1 Blocker
+1 blocker, also I once again cash in my "we really shouldn't have two separate custom partitioning UIs" card!
That's enough votes to accept this.
FEDORA-2020-efbff266dc has been submitted as an update to Fedora 32. https://bodhi.fedoraproject.org/updates/FEDORA-2020-efbff266dc
(In reply to Fedora Update System from comment #22) > FEDORA-2020-efbff266dc has been submitted as an update to Fedora 32. > https://bodhi.fedoraproject.org/updates/FEDORA-2020-efbff266dc Verified fixed with RC 1.5.
FEDORA-2020-efbff266dc has been pushed to the Fedora 32 stable repository. If problem still persists, please make note of it in this bug report.