Hi,

One morning after running Arch Linux fairly problem-free on this PC (a laptop) for six months or more, it wouldn't boot due to several systemd services failing on startup. Some related to partitions (including my swap partition) not mounting properly, it seems like all partitions listed in /etc/fstab are affected, including my root partition and at least one other systemd service fails (the journal flush one). What's more I found my other bleeding-edge distributions using the 4.15.3 or later kernel and systemd (e.g. Debian sid, Fedora 27, openSUSE Tumbleweed) weren't booting due to similar issues.

I tried looking at /etc/fstab (which, for some distributions, used disk locations, instead of using UUID to identify them, which can change, of course, so I know that's not the issue), I tried running fsck and btrfs check to check for file system errors (of which there were none) and I was even wondering if this was a Btrfs issue and to see I did a clean install of Arch with the ext4 file system and the problem persisted.

There was just one thing I found that worked, reverting to an older kernel. The 4.15.3 and earlier kernels worked fine.

Gentoo (with OpenRC instead of systemd), with the newer kernels, boots fine, but then it doesn't take long before no commands (including those used to launch programs like xterm) work (i.e. you can execute the ls command and it just hangs, no output is returned and the command doesn't exit). Void (with runit), with the newer kernels, suffers from the same affliction too.

I have never experienced any issues like this before.

I should mention I first noticed this error approximately a fortnight ago, I was hoping that in a few kernel releases this issue would have been rectified. Sadly this is not the case, even with the 4.15.7 kernel this issue persists.

If you want to check out a fstab from a system with a problem here's my Arch fstab:

# Static information about the filesystems.
# See fstab(5) for details.

# <file system> <dir> <type> <options> <dump> <pass>
/dev/sda1 /boot/efi vfat defaults 0 1
/dev/sda2 swap swap defaults 0 0
/dev/sda3 / ext4 defaults 0 1
/dev/sdb1 /data ext4 defaults 0 1



I am not a newbie (been using Linux for six years this year, I've successfully installed around 80 different operating systems including Exherbo and LFS, so I'm confident I'm not what you'd call a 'newbie'), so I do not think this belongs in the newbie forum, nor does it belong in the hardware forums as this seems to more relate to software, as for the software forum well to me it seemed more like it was for people with application software issues, hence why I'm asking here.

Here's (https://imgur.com/xV1Aobc.png) a screenshot of what happens when I try to boot Arch Linux on ext4, with the 4.15.7 kernel and systemd 237.

I doubt it'd help but here (http://paste.ubuntu.com/p/7FhDm3ckXH) is the output of lspci on my system.

Thanks for your time

Try adding "mount.timout=300s" to your grub entry. See if the devices appear then.

I was wrong it wasn't the root partition that wasn't being mounted it was the other partitions, like the swap, EFI and data (ext4) partitions that weren't. Adding that to the linux line in the GRUB entry didn't seem to change the times, they were still 1 & 1/2 minutes (90s). Guessing it'll only work for the root partition.

Shouldn't - there is a specific rootdelay for the root.
This is systemd not waiting long enough for udev to create the device nodes I think. If it were me I'd try UUID (which I don't usually use) or remove all those entries from fstab (just comment them out). See what happens. Remove any resume parms from grub if using swap for it.

Forgot to mention, I'm currently on 4.15.6 no issues - different chipsets of course.

Well changing to UUIDs got me further, just /dev/sda2 (swap) failed to mount (when I didn't comment out anything). Oddly commenting out /dev/sda2 from fstab failed to stop udev from trying to mount it. That's even when sda2 is represented by UUID. Also tried running `swapoff /dev/sda2` from a chroot into the Arch system (with dev, sys rbinded and proc mounted with mount -t). When I used UUID (without commenting out) these are the most informative screenshots I got:

https://imgur.com/K4eWXsL.jpg
https://imgur.com/iKPnX6a.jpg

(sorry my phone is really old, so pics are not high-def)

The result was similar with commenting out those entries.

These issues went on for at least 5 minutes (i.e. at the point I got the idea that it wasn't going to boot and I rebooted).

Guessing you're completely stuck on what I should do?

I'm not
just install slackware
sure to solve ALL of your systemd problems