During installation process I setup Fedora BTRFS subvolumes in order to proper work with Timeshift software (@ and @home), but I saw that while creating a Snapshot it takes ~60s while for Arch/Manjaro it’s instantaneous like 2s.
I can confirm this long time running Fedora and also Nobara (Fedora based) on virtual machine and also running on the hardware.
Is there any particular reason for Timeshift to be slower on Fedora?
Note: I also checked OpenSUSE Tumbleweed with Snapper and it create Snapshot a lot faster too.
Once my knowledge on Linux is limited in general I always use default options, with is the case for BTRFS.
Could you share the command line to check quota status, so I can check and post here the result to see if this is the reason for the performance difference.
The longer time is spent at the phase:
preparing… (gear spinning)
Deleting snapshot is slow as creating new one, it takes a lot of seconds at the same phase:
preparing… (gear spinning)
On fedora I just change subvolumes names for @ and @home and that is all.
On Nobara it’s done by default once it uses a different installation (calamares)
I don’t know if there’s any workaround besides modifying the source for timeshift and removing the notify-send command. I’m not sure why other distros don’t have this issue. I use Fedora KDE, and judging by your screenshots it seems like you do also. Maybe it’s a Fedora KDE related issue with timeshift? Do you use wayland?
Nice you were able to find the issue, it may worth to open a bug report on Timeshift github once you have some data. Unfortunately I don’t have such know how fix the code myself. Thanks god there so many wise developers out there,
I’m using KDE too on Manjaro, I’ve been playing around on Virtual Machine to learn more how to use fedora and fedora base distros. I have a new desktop build and ready with nobara installed just waiting HD swap with personal data to move on definitive. Just preparing a smooth transition.