You are not logged in.
I just did a system update
pacman -Syu Then a quick restart later I cannot start i3 because
$DISPLAY and
no Screen found errors. (Those aren't the exact errors I saw, just what I remember).
I also used a live usb to chroot into my installation and try updating again thinking if something got messed up during the update process, this could fix it. But I was wrong ![]()
systemctl status systemd-modules-load.service shows this msg
systemd[1]: Starting Load Kernel Modules ...
systemd-modules-load[602]: Failed to lookup module alias 'crypto_user': Function not implemented
systemd[1]: systemd-modules-load.service: Main process exited, code=exited, status=1/FAILURE
systemd[1]: systemd-modules-load.service: Failed with result 'exit-code'.
systemd[1]: Failed to start Load Kernel Modules. grep boot /proc/self/mounts shows this
/dev/nvme0n1p5 /boot ext2 rw,relatime,block_validity,barrier,user_xattr,acl 0 0 modprobe dm_mod shows this
modprobe: FATAL: Module dm_mod not found in directory /lib/modules/4.20.0-arch1-1-ARCH Although looking at the repos
linux 4.20.1.arch1-1 seems to be the latest version. I don't remember which one got installed when I updated the system but I think it should be the one in the repo. So why would kernel version show 4.20.0-arch1-1? I suspect this could be the thing that's causing the error.
Last edited by atis (2019-01-15 12:58:08)
Offline
What's your output for
pacman -Q linux
uname -a the versions reported there have to match
Did you have that /boot partition mounted while doing the update? If not then that's likely to be your issue
Offline
pacman -Q linux shows
linux 4.20.1.arch1-1 and
uname -a shows
Linux atis 4.20.0-arch1-1-ARCH #1 SMP PREEMPT Mon Dec 24 03:00:40 UTC 2018 x86_64 GNU/Linux It is (was) as working system. I was just doing a routine weekly update at the time.
I assume if you booted into a working system, /boot should be mounted by default, should it not?
Offline
I just tried updating from live-usb with EFI partition mounted on /boot.
mkinitcpio -p linux does show
Starting build: 4.20.1-arch1-1-ARCH but after it's done,
pacman -Q linux and
uname -a still show different versions.
Offline
Just for clarification the uname -a should be run from the actual system not the chroot. /boot doesn't have to be mounted for a successful boot it depends on what is configured in your boot loader. if that one points to the /boot partition then it will boot from there, however if it points to the /boot directory on your root partition, it will boot from there. Depending on which of the two is the case, you are installing the kernel to the wrong location (if /boot as in the partition is unmounted the kernel will be written to the /boot directory on your root filesystem, and not to the partition)
Offline
Both
arch-chroot and the actual system report same output on
uname -a . I should add I'm using systemd-boot not grub.
if that one points to the /boot partition then it will boot from there, however if it points to the /boot directory on your root partition, it will boot from there
That could be it. At first I did the normal install with
/boot mounted from root's. Just now I ran
pacman -S linux with my EFI partition on /boot path (Windows' EFI partition. It's a dual boot machine).
Both of the times it showed different results on either of those commands.
EDIT: /boot partition mentioned in my fstab is the one loaded in root. Windows' EFI isn't mentioned in it.
Last edited by atis (2019-01-14 11:56:23)
Offline
Right now I have downgraded the kernel and things seems to be fine now.
This is what
lsblk shows
NAME MAJ:MIN RM SIZE RO TYPE MOUNTPOINT
nvme0n1 259:0 0 477G 0 disk
├─nvme0n1p1 259:1 0 499M 0 part
├─nvme0n1p2 259:2 0 100M 0 part
├─nvme0n1p3 259:3 0 16M 0 part
├─nvme0n1p4 259:4 0 237.9G 0 part
├─nvme0n1p5 259:5 0 400M 0 part /boot
└─nvme0n1p6 259:6 0 238.1G 0 part
└─lvm 254:0 0 238.1G 0 crypt
├─volgroup0-lv_root 254:1 0 30G 0 lvm /
└─volgroup0-lv_home 254:2 0 208.1G 0 lvm /homeOffline
You really have to figure out what's what here. If you use systemd-boot the case is pretty clear, unless you chainload another bootloader your /boot partition and the location of your linux images must (well technically, no, but you save yourself a lot of unnecessary scripting and additional work if that is the case) be the same as the ESP (so the ext2 partition in your OP cannot be correct, it should be a FAT variant).
Last edited by V1del (2019-01-14 16:15:54)
Offline
Ah I see! So while
nvme0n1p5 gets mounted at
/boot it isn't actually doing the job. I think I made that one when I thought I didn't have windows already installed but then I found out, so I installed systemd-boot on the ESP partition (nvme0n1p2) but the old
/boot remained in my
/etc/fstab I think I know what to do. Boot into
arch-chroot via Live-USB. Mount proper partitions at proper locations, so first root, then boot (nvme0n1p2) and then home.
rm -rf that old fstab. Generate a new one and then run
pacman -Syu which hopefully should fix everything.
Thanks for the help! I'll try this asap and report back.
Last edited by atis (2019-01-14 17:06:02)
Offline
Hey V1del, that was the issue and I did exactly as my post above and it worked! Everything is working without a problem. Thanks for the help!
Offline
Glad to hear, yes, this can be a bit confusing if you are used to BIOS boots and might've used a bootloader that can read from linux filesystems. But in this case you have to ensure the images are written to the EFI partition for them to be read by systemd-boot
Offline