Extremely slow boot

Hi,
I do have a Arch specific boot problem… Hardware is a Lenovo B550 laptop 4GB RAM 512GB SDD, newest kernel (but issue is present with lts-kernel, too). It does not matter if I boot installed system or live system via USB.System hangs about 1.5 minutes after loading initial ramdisk. Last I see on screen is EDD probing (which seems to run successfully) - then there is no output on the screen for ~1.5 minutes - then booting finishes rapidly. ‘systemd-analyze time’ tells me 15 seconds - which is the time after the “hang” and it is perfectly. But how can I debug the hanging BEFORE systemd gets control? I already started with kernel parameter loglevel=7 but there is no output that is showing any trouble. Starting of other Linux systems (all older kernels) goes very fast. So I ned a debug info between “kernel load” and “systemd takes system over”…

Does the system work from a TTY?
Press Ctrl-Alt-F2, you should get a full screen terminal where you can give normal terminal commands (but it doesn’t support GUI programs), e.g. create logs for us.

No, it does not work from a terminal, too. But I have stepped further: giving acpi=off as grub kernel parameter gives normal boot speed - but of course nothing that is acpi-related does work. acpi=ht throws system back to extremely long boot time. Other distros work without disabling acpi. I do have the same issue on three totally different machines - and on tons of others not… So I think I must do some “acpi debugging”… No idea how I can do that.

Can you show the tech specs (a link?) of the system?

https://mobilespecs.net/laptop/Lenovo/Lenovo_B550.html

Thanks. It has Nvidia G210M graphics card. So one thing to try is not use the nvidia driver but nouveau instead.
Did you boot the ISO with the default option of the boot menu, or with the Nvidia option?

journalctl -b -0

boot log would give a hint for sure…

1 Like

I boot with the default option. Bootlog shows nothing interesting because of it starts when systemd takes control. Delay is BEFORE systemd takes control. Delay is about 90 seconds - booting after delay (from greeting “welcome to endeavouros…” over all starting daemons and services only takes 15 seconds. But of course waiting 90 seconds without seeing any messages or finding anything in the logs (because the logs do start AFTER the delay) is not amazing… So I need a solution for getting debug infos and gathering them into a log from the time the kernel wakes up until systemd takes over control.

Maybe @joekamprad knows if the ISO supports that somehow.

But you might also try the Arch ISO (https://www.archlinux.org/download) and see if you can boot with it. If so, then you could get the logs. Of course it is not the same, but may still give some more info about the case.

I tried booting with Arch ISO. Delay is same (about 90 seconds). journalctl -b -0 tells boot time of 10 seconds. Delay is BEFORE log starts.

Maybe try sudo journalctl -p 3 -xb ?

EDIT

Also return sudo systemd-analyze blame please. :+1:

2 Likes

It is nvidia discrete graphics so 340xx driver. Nouveau is option.

After kernelupdates yesterday delay is much shorter - but always present (standard kernel and lts kernel) I set loglevel to 7 as grub kernel parameter and attach three logs:
journalct -b 0 : boot.txt
journalctl -p 3 -xb: jctl.txt
systemd-analyze blame: blame.txt
ED: Because of simplest file format (.txt) is not allowed to attach I renamed (!) all three to pdf.

Same SSD inserted in another computer with slower CPU boots in 15 seconds. The “wait state” before systemd takes over is about 15 seconds now (former 90 seconds).blame.pdf (3.6 KB) boot.pdf (152.8 KB) jctl.pdf (337 Bytes)

your links do not work for me. Is it a problem to just copy and paste the output from terminal here with three ``` before and after the output? Besides the fact we don’t know you and are not interested in downloading some random pdf file.

hi friend, open with libreoffice

1 Like

I tried that already, but not possible here. Message from forum soft:

Body is limited to 32000 characters; you entered 160610.

1 Like

Use any simple text editor for opening. It is “plain text” format - the best to cat and grep inside.

1 Like

Ok got it. Sorry not at a linux box atm. This looks like a 10year old all in one machine correct? If that’s the case the first thing i would start with would be to update the bios if possible. It looks to be the one the machine shipped with new.

It is a laptop. I have looked around much but I cannot find any BIOS download for it. That was the first thing I have had started…

The kernel log levels are:

0 (KERN_EMERG)

The system is unusable.

1 (KERN_ALERT)

Actions that must be taken care of immediately.

2 (KERN_CRIT)

Critical conditions.

3 (KERN_ERR)

Non-critical error conditions.

4 (KERN_WARNING)

Warning conditions that should be taken care of.

5 (KERN_NOTICE)

Normal, but significant events.

6 (KERN_INFO)

Informational messages that require no action.

7 (KERN_DEBUG)

Kernel debugging messages, output by the kernel if the developer enabled debugging at compile time.

I personally use loglevel=3