archinstall
archinstall copied to clipboard
Crash in attempting to do arch install with manual partitioning
Which ISO version are you using?
2024-04-01
The installation log
https://0x0.st/8_-3.log
describe the problem
Description of the issue
I was installing on a ThinkPad T480.
Then I went through standard install steps, both with manual ext4 and btrfs file system setups. No matter what I try, it crashes!
So there's a couple of attempts in this log, and at least one of them is trying to re-use existing partitions, and also create new subvolumes at the same time. Causing it to crash because @home already exists. So new subvolumes should not be created when re-using partitions.
Except I am not reusing partitions, I attempt to delete first and recreate and the same crash comes. In general, the crash appears.
I'm having the same issue on qemu
The March ISO hadn't this issue, if I remember correctly
BTW, manual partitioning related issues are recurrent bugs on Archinstall ,that appear time by time.
I think the team should consider to hardly revisit the way that archistall manage manual paritioning , and do more tests about before pushing updates
i experienced the same issues, however i found a workaround
i used gparted to delete partitions (to wipe them), hoping that if the partitions were already made then the issue would go away if archinstall just created a new partition with free space, and it did
so it seems to be triggered by wiping an existing partition at least (don't know about deleting an existing one, can't remember if i tried that or not. i was frustrated (nvidia broke my system again)).
as long as your installing onto "free space" then it should be fine... hopefully
Maybe related: https://github.com/archlinux/archinstall/issues/3689
In the presence of an existing BTRFS (which must be kept as it is, and must not be mounted), there's no way to install...
Of course, this makes the installer completely unusable (I want to have several installations on my machines) and I agree with @thealio
I think the team should consider to hardly revisit the way that archistall manage manual paritioning , and do more tests about before pushing updates