First, with 38:
1 - did a routine dnf upgrade. This corrupted the partition and required booting from the live usb and using fsck with a backup superblock.
2 - Rebooted 38 and ran the Software app. Clicked the download button to get 39. After it finished, nothing happened.
3 - upgraded workstation using the DNF system plugin
Then, on 39:
1 - 8 of my desktop icons were greyed out with locks on them.
2 - some, I was able to use the popup to mark executable.
3 - Four of them were symlinks to elsewhere and I had to use chmod +x to mark the targets as executable before they would run in the user account.
All in all, the process wasted about 2 hours - not including the 6 hours it took to download and install the 5.1 GB of the 39 upgrade.
I would be reluctant to suggest this distro to anyone interested in switching from Mac or Windows. Debian 12 and Linux Mint 21 are both much faster installs and more seamless, imho.
If I had realized how much time and effort would be required, I might have decided to stay on 38.
The issue that occurred with the upgrade of F38 in step 1 was a red flag that should have immediately halted any further updates until the cause of that problem was resolved.
I am unsure that I understand why you are moving back and forth between dnf and the gnome software app.
In step 1 you said you used dnf and did a routine dnf upgrade which appears to have had an issue with file sytem corruption.
In step 2 you then used the software app. At that point nothing was said about repeating the dnf upgrade with the always recommended dnf upgrade --refresh command to 100% confirm that the system is fully in sync with the latest in the repo before continuing a release version upgrade. To me this also is a red flag because it would appear the earlier upgrade may not have properly completed and should have been verified fully before continuing.
Finally, in step 3 you said you used the plugin but provided no details. It always helps us to know exactly what was done by posting the exact steps followed. The dnf system-upgrade process has 2 main parts. ‘Download’ which provides another reminder to perform the dnf upgrade --refresh step before continuing. Then after the successful download completes that is followed by the ‘Reboot’ step to perform the actual upgrade.
I have always used the dnf commands from the command line since I know that there are sometimes inconsistencies in switching back and forth between the 2 package managers (dnf on the cli and the gnome software gui). Since I never switch back and forth I have never experienced problems of the sort you describe.
The issue that occurred with the upgrade of F38 in step 1 was a red
flag that should have immediately halted any further updates until the
cause of that problem was resolved.
I am unsure that I understand why you are moving back and forth
between dnf and the gnome software app.
I’m not. I normally use dnf so Id first did my normal dnf upgrade on 38.
I noticed that there was a kernel update was included so i thought I had
better reboot before continuing. That’s the step that zapped the
partition. After I got that repaired I the followed the instructions on:
It recommended the gnome software app (I don’t like anything about
gnome) so I used it. As I said, it downloaded something and then did
nothing so I the followed the instructions on:
Correct. The reboot was appropriate, but may have been done before the upgrade fully completed. If the upgrade was still writing to the drive a reboot performed too soon may have caused the corruption.
I tend to disagree with the recommendation to use the gnome software app for performing release upgrades. It hides everything done within the gui so the user has no control at any stage.
Ever since it was made available I have used the dnf system-upgrade process for release version updates.