Error when trying to boot up Arch ISO from USB drive: invalid magic

FRVNT · 2024-03-04 19:25:02

Progress update: VERSION 2023-08-01 (kernel 6.4.7) BOOTS SUCCESSFULLY
I will bissect until I find the latest working version

Last edited by FRVNT (2024-03-04 19:31:59)

seth · 2024-03-04 19:33:50

From the other thread, isos from mid 2023 should still work.
Edit: ok, you found that one

@cryptearth, i'll again point out that 16 months old raptor lakes are affected.
If you post this unfounded bullshit in one single more thread, I'll ask for you to be banned for trolling.
I'm not saying this again.

Last edited by seth (2024-03-04 19:34:35)

FRVNT · 2024-03-04 19:35:44

So should I label this as solved?

V1del · 2024-03-04 19:42:50

@cryptearth you're talking verifiable incorrect statements based on feelings, seth linked threads with 2022 hw having issues, i can link you threads with the same hardware op has still having an actually working Arch install - where a bug in a particular library had an issue with the old HW indeed, but that was a bug that was fixed -- for the old HW. https://bbs.archlinux.org/viewtopic.php?id=292236

From an "instructions" perspective AVX2 was the last bigger change on x86-64, 10 years ago. But this is still not generally speaking the assumed baseline of the distribution.

EDIT: F5 and even 20mins ago...

Last edited by V1del (2024-03-04 19:45:41)

FRVNT · 2024-03-04 19:44:10

Tried version 6.4.7, it didn't boot but the error message might be of use to you [ https://ibb.co/L9m6DTK ]

FRVNT · 2024-03-04 20:22:27

Version 6.2.8 boots as well, looks like I will stick to this.

Last edited by FRVNT (2024-03-04 20:27:25)

seth · 2024-03-04 21:28:36

You might be interested in https://aur.archlinux.org/packages/linux-lts61
So wrt #26 and #30 the behavior turned non-deterministic with 6.4.7 for you?

Did you try any versions inbetween 6.2.8 and 6.4.7?
Was this all with install isos or also on an installed system?

The error in #30 is a follow up, the iso could not be properly decompressed, you're lacking modules in the initramfs and can't access/mount the root FS)

duaner · 2024-03-05 02:24:09

cryptearth wrote:

Well - get yourself such old hardware and try it yourself - or something like gentoo or go extreme with LFS - you'll see the hardware is just too outdated - it just lacks required physical implementations of modern instructions.

That's funny. I'm still running arch (about ten years now?) on an intel core2 duo system built close to twenty years ago. I did break down and update it to 8G ram (from 2G) about five years ago, but that was more to support firefox than anything. Silverstone CW01 (boat anchor) case, asus p5q-em motherboard, etc. My octogenarian father uses it daily. It's so little trouble to maintain that I tend to forget about it, but I update it every month.

Last edited by duaner (2024-03-05 02:26:08)

puzzle · 2024-03-16 13:43:59

seth wrote:

There's a recent-ish pattern of failures to decompress the initramfs. Limiting the amount of cores has been reliably successful in some cases (leading to the assumption it might be because of the multithreaded zstd decompress) but hasn't worked in several others.
At least one user was able to decompress images <= ~15MB but not larger ones - w/ the 6.7 kernels (no such problem w/ the 6.6 kernels)
However the general problem predates the 6.7 release. It's weird.
It's possibly a race condition and limiting the cores might have just slowed down the affected systems so much that they escaped the race.
The reason for the behavior isn't understood, but it is FOR SURE not the age of your hardware design.
Eg. https://bbs.archlinux.org/viewtopic.php?id=290093 and https://bbs.archlinux.org/viewtopic.php?id=291171 are raptor lakes from 2022.
As you can see there, the behavior isn't even deterministic (on an installed system) - the relation between CPU and IO speed might be a factor
==> Incidentally: tried a different USB port?
https://github.com/torvalds/linux/blob/ … compress.c hasn't changed in 4 years and nothing in https://github.com/torvalds/linux/commi … nitramfs.c looks even remotely like a possible culprit.

------------
@cryptearth, just as a reminder: https://bbs.archlinux.org/viewtopic.php … 6#p2151646

~~After I upgraded linux-lts from 6.6.18-1 to 6.6.21-1, the same error occurred when booting with grub.~~

If I boot from Grub Shell,I can enter the system.

grub> linux /boot/vmlinuz-linux-lts root=/dev/sda1 rw
grub> initrd /boot/initramfs-linux-lts.img
grub> boot

I found out that the problem was caused by microcode
because upgraded mkinitcpio (38-4 -> 38.1-1)
After following this post, everything is now normal
https://bbs.archlinux.org/viewtopic.php?id=293437

Last edited by puzzle (2024-03-16 15:24:24)

seth · 2024-03-16 14:49:06

If I use the grub command number linux initred boot

Errr… can we have the english version of that?

Maybe relevant: https://bbs.archlinux.org/viewtopic.php?id=292840 hints at BIOS vs. UEFI cause

Arch Linux

#26 2024-03-04 19:25:02

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#27 2024-03-04 19:33:50

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#28 2024-03-04 19:35:44

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#29 2024-03-04 19:42:50

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#30 2024-03-04 19:44:10

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#31 2024-03-04 20:22:27

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#32 2024-03-04 21:28:36

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#33 2024-03-05 02:24:09

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#34 2024-03-16 13:43:59

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

#35 2024-03-16 14:49:06

Re: Error when trying to boot up Arch ISO from USB drive: invalid magic

Board footer