Bug 29708

Summary:	Mageia Grub can't recognize kernels after update
Product:	Mageia	Reporter:	Sabungayam 8 <Sabungayam8>
Component:	RPM Packages	Assignee:	Kernel and Drivers maintainers <kernel>
Status:	RESOLVED INVALID	QA Contact:
Severity:	critical
Priority:	Normal	CC:	doktor5000, ftg, marja11
Version:	8
Target Milestone:	---
Hardware:	x86_64
OS:	Linux
URL:	https://forums.mageia.org/en/viewtopic.php?f=7&t=14396
Whiteboard:
Source RPM:		CVE:
Status comment:

Description Sabungayam 8 2021-11-28 07:13:39 CET

Description of problem:
With recent updates 5.10.75-desktop-1.mga8, 5.10.78-desktop-1.mga8 and 5.15.4-desktop-1.mga8 I have ended up with a message:

[code]

[ 0.978687] pci 0000:00:07.0: DPC: RP PIO log size 0 is invalid
/init: line 153: /user/lib/systemd-udevd: Too many levels of symbolic links
failed to send request to update environment: No such file or directory

[/code]

I have now 3 machines with this problem.

Version-Release number of selected component (if applicable):

How reproducible:
I started with one machine and it was bootable but this was a dualboot with another systems Grub in charge. In the forum help was gotten on how to direct
chainload to Mageia's grub only to get the same message.

I later updated my work PC with Mageia on it (I think it was Thursday or Wednesday). This machine has Mageia grub in charge and also properly in EFI. I was informed that Mageia grub has issues if
booted only from partition.

I just updated today another PC AMD. This install was pretty fresh and a net-install. The kernel there was 5.10.62-desktop-1.mga8. I updated it to 5.15.4-desktop-1.mga8. Same message and there is NOT a kernel that I can boot. This is dual boot so Fortunately I can tell it to boot into Debian but can't even load any Mageia kernels.

I am on a Mageia machine now but I actually installed Refind here, but I am still reluctant to currently update it and run into any more issues.

I have one more machine which can be tried as well if more info is needed. This will be like my work Machine, Mageia only. (It has not been updated recently and I know it will have older kernel. Maybe 5.10.62-desktop-1.mga8.

I am able to boot two machines. removing kernels does not change anything. On one machine I am able only to boot 5.10.62-desktop-1.mga8. (I had been booting 5.10.70-desktop-1.mga8 as the top kernel until this issue began) I think I can boot 5.10.70-desktop-1.mga8 on my work machine. I do know I had to drop back about 3 kernels on my work machine before anything was booting. So it may be 5.10.62-desktop-1.mga8 or lower.

This is with Mageia grub.

Steps to Reproduce:
1. Have a Mageia 8 install sitting around maybe kernel 5.10.70-desktop-1.mga8 or 5.10.70-desktop-1.mga8 (Seems to be where my problems began.)

2. Just do a normal update using Mageia updater.

3. Try to reboot and get that message constantly until maybe you can backtrack to a kernel that it doesn't give that message on. If you haven't enough kernels maybe end up being unable to boot into Mageia.

Have tried to be as concise as possible. There is a forum posting on this issue with the same title as this bug.

Thank you.

Comment 1 Frank Griffin 2021-11-29 16:23:43 CET

Assigning to the kernel group, as I'm guessing grub comes under that.

Assignee: bugsquad => kernel
CC: (none) => ftg

Comment 2 Marja Van Waes 2021-11-29 22:36:06 CET

(In reply to Sabungayam 8 from comment #0)

<snip>
> 
> Have tried to be as concise as possible. There is a forum posting on this
> issue with the same title as this bug. 
> 

And here is the link:
https://forums.mageia.org/en/viewtopic.php?f=7&t=14396

Do you also have this issue on a machine that has no more than one Linux installation on it (so only one Mageia and no other other Linux distribution)?

URL: (none) => https://forums.mageia.org/en/viewtopic.php?f=7&t=14396
CC: (none) => marja11

Comment 3 Sabungayam 8 2021-12-08 07:48:14 CET

(In reply to Marja Van Waes from comment #2)
> (In reply to Sabungayam 8 from comment #0)
> 
> <snip>
> > 
> > Have tried to be as concise as possible. There is a forum posting on this
> > issue with the same title as this bug. 
> > 
> 
> And here is the link:
> https://forums.mageia.org/en/viewtopic.php?f=7&t=14396
> 
> Do you also have this issue on a machine that has no more than one Linux
> installation on it (so only one Mageia and no other other Linux
> distribution)?

Sorry, I missed you reply. Here, is the current situation. 

1st to answer your question, yes, I have a Mageia that isn't dual booted and it was suggested to troubleshoot the issue there so as to avoid confusing other possible issues. 

That has been looked at in this thread here: 

https://forums.mageia.org/en/viewtopic.php?f=7&t=14405&p=84562#p84562

Comment 4 Sabungayam 8 2021-12-08 07:52:52 CET

https://forums.mageia.org/en/viewtopic.php?f=7&t=14405&p=84562#p84562

In this thread what has been discovered is this:

The problem was caused by this VPN Astrill, the library "liblsp.so" from Astrill interferes with the dracut program and screws up every machine. 
 
Removing the VPN helps to rectify the problem.


I had recently installed this VPN RPM on my Mageia installs. 

To reproduce, just install Astrill RPM from their site. 

With Astrill removed, performed an update with no problem at all. Now running current kernel from latest update today.

Comment 5 Florian Hubold 2021-12-08 23:30:01 CET

Closing as invalid as dracut works just fine to recreate the initrd without the 3rd party VPN client.

Status: NEW => RESOLVED
Resolution: (none) => INVALID
CC: (none) => doktor5000