Well, this is weird…
After two years of a happy honey moon with Manjaro, I got a read only root partition and a kernel panic after a reboot on my daily driver Thinkpad T440s.
Booting a live USB was a breeze, but then I missed the good ol’ days of Debian or even Red Hat / CentOS text based install tools: the Manjaro installer (Calamare’s the name…?) hangs when trying to create a new Volume Group. No matter if I got the Plasma live USB image or the Deepin one –this is what I really want. It hangs at the very same step.
What do I have:
Two physical SSD disks: one with the OS-that-runs-better-with-windows (license already payed, so…) with a minimal free space, a
/bootext4 partition and a free space available as a Physical Volume for my say gabrielvg Volume Group.
Eventually, I bought another beautiful NVMe SSD disk and created another PV, so the VG was extended.
Now, a more deep step:
The Volume Group was populated with three Logical Volumes:
rootlv, which holds my BTRFS
swaplv (guess what was in there…), and a password-at-boot-time-proyected crypto device
/home partition. Honestly, I don’t remember if it’s LUKS. May be so.
I do have copies of my
/etc/mkinitcpio.conf in a safe place.
What do I want:
Yes, you are thinking inside the box: I want Manjaro with Deepin DE and I do want to keep my cypher
/homepartition unchanged and accessible as before the kernel panic situation. So, I thought: «Lets rebuild the volume group from zero!». I wiped the
/rootpartition, deleted the file system, then the partition and even the logical volume. Said bye-bye to my swap, too. Just in case.
A fresh start with only an encrypted item. Well, let’s go!
Some fancy housekeeping after booting the live minimal Manjaro Deeping and launched the installer for the fourth time.
Selected the main disk (the one with MS-Windows and the the other Linux stuff) and selected the little
/bootpartition. Checked it for being formatted and to be mounted as
/boot. So far, so good.
Didn’t touch the Windows partitions by now. Not important in this stage.
Moved into the second “device” marked as the ‘vgroot’, the name I used to use with the VG before the crash. And there it is the
/homeencrypted partition and the rest of the disk, free (remember I deleted all the remains of my latter system, R.I.P.).
Now, I create a new partition and mark it as ‘LVM PV’, so I can create a new VG. When I click on ‘CREATE VG’, Calamares do some checks and hangs. No error messages, no console output
demsglogging a segmentation fault:
[ 4525.430142] calamares: segfault at 7f224e5bbe50 ip 00007f25deede20e sp 00007ffea4d23f60 error 4 in libkpmcore.so.21.08.2[7f25deead000+9f000]
So, I think: «I need to work it out it in the CLI». I know how to manage LVM in Linux (or IBM AIX) using the CLI but the real thing here is that is the installer who needs to manage the LVM set up, not me from the outside. Am I wrong?
I am stuck here, really thought It will be easier for a modern distro to solve an easy situation like that.
Any hints would be highly appreciated.
Thanks in advance, folks