Bug 24731

Summary:	segfaults when following wiki upgrade instructions
Product:	Mageia	Reporter:	Remco Rijnders <remco>
Component:	RPM Packages	Assignee:	Mageia Bug Squad <bugsquad>
Status:	RESOLVED DUPLICATE	QA Contact:
Severity:	normal
Priority:	Normal	CC:	marja11, tmb
Version:	7	Keywords:	NEEDINFO
Target Milestone:	---
Hardware:	x86_64
OS:	Linux
Whiteboard:
Source RPM:	?	CVE:
Status comment:
Attachments:	logfile failed upgrade

Description Remco Rijnders 2019-04-28 19:29:03 CEST Comment hidden (obsolete)

Description of problem:
Today I updated my Mageia 6 installation to 7. During the upgrade the installation of the kernel fails with the error:

Installation failed: microcode < 0.20180807-4 conflicts with kernel-desktop-5.1.0-0.rc6.2.mga7-1-1.mga7.x86_64

I found out that I had both microcode version 0.20180807 packages 1.mga6.nonfree as well as 4.mga7.nonfree installed on my machine. After manually removing the mga6 package, I was able to install the kernel.

Marja Van Waes 2019-04-29 08:48:27 CEST

Assignee: bugsquad => kernel
CC: (none) => marja11

Marja Van Waes 2019-04-29 08:48:39 CEST

CC: (none) => tmb

Comment 1 Thomas Backlund 2019-04-29 23:44:44 CEST

The conflict is in place to force that the newest microcode is installed before the kernel so it gets included in the initrd for the new kernel as we rely on early microcode loading...

The real question is how you managed to get 2 microcode packagees installed ?
Did you abort the update at some point in mid of package updates ?

Comment 2 Remco Rijnders 2019-05-05 15:04:50 CEST

(In reply to Thomas Backlund from comment #1)
> The conflict is in place to force that the newest microcode is installed
> before the kernel so it gets included in the initrd for the new kernel as we
> rely on early microcode loading...
> 
> The real question is how you managed to get 2 microcode packagees installed ?
> Did you abort the update at some point in mid of package updates ?

Ah, funny that you should mention that. I followed the upgrade instructions at https://wiki.mageia.org/en/Mageia_7_Release_Notes#Upgrading_online.2C_using_urpmi_.28CLI.29 and about 200 packages in the progress urpmi (or one of the underlying tools used) segfaulted. This happened repeatedly till I rebooted and the problem then went away.

So, it seems that what I observed with the microcode packages conflicting was a symptom of another problem and not a problem with the microcode packages themselves. As I don't think I took notes of the segfault issue, I am afraid I have little additional information to share on that.

Comment 3 Marja Van Waes 2019-05-05 15:40:14 CEST

(In reply to Remco Rijnders from comment #2)
> (In reply to Thomas Backlund from comment #1)
> > The conflict is in place to force that the newest microcode is installed
> > before the kernel so it gets included in the initrd for the new kernel as we
> > rely on early microcode loading...
> > 
> > The real question is how you managed to get 2 microcode packagees installed ?
> > Did you abort the update at some point in mid of package updates ?
> 
> Ah, funny that you should mention that. I followed the upgrade instructions
> at
> https://wiki.mageia.org/en/Mageia_7_Release_Notes#Upgrading_online.
> 2C_using_urpmi_.28CLI.29 and about 200 packages in the progress urpmi (or
> one of the underlying tools used) segfaulted. This happened repeatedly till
> I rebooted and the problem then went away.
> 

If you remember in between which days or times that happened, then please  attach log.txt that is the result of running, as root:

  journalctl -a --since="YYYY-MM-DD hh:mm" --until="YYYY-MM-DD hh:mm" > log.txt

and adjust the --since time to right before you started to follow the wiki upgrade instuctions and the --until time to when you rebooted after the repeated segfaults.

Summary: Mageia 6 to 7 upgrade fails due to microcode package => segfaults when following wiki upgrade instructions
Keywords: (none) => NEEDINFO
Source RPM: microcode => ?
Assignee: kernel => bugsquad

Comment 4 Remco Rijnders 2019-05-05 16:23:56 CEST

Created attachment 10979 [details]
logfile failed upgrade

Thanks Marja, please find attached the requested part of the logfile.

Comment 5 Marja Van Waes 2019-05-05 16:56:54 CEST

Thanks, Remmy :-)


urpmi was called 4 times, after the 3rd time there was a trap divide error

Apr 28 11:26:40 ... urpmi[2602]: called with: --auto-update

Apr 28 11:32:51 ... urpmi[29637]: called with: --replacefiles --auto-update --auto

Apr 28 11:40:11 ... urpmi[29637]: called with: --previous-priority-upgrade=rpm,perl-URPM,perl-MDV-Distribconf,urpmi,meta-task,glibc,aria2 --replacefiles --auto-update --auto

    Apr 28 11:49:31 ... kernel: traps: urpmi[29637] trap divide error
    ip:7f33ae156876 sp:7ffd0d4eaae0 error:0 in libdb-5.3.so (deleted)
    [7f33ae082000+1b1000]            ==That is bug 24767==

Apr 28 11:49:57 ... urpmi[8587]: called with: --replacefiles --auto-update --auto

    Apr 28 11:50:17 ... kernel: traps: urpmi[8587] general
    protection ip:7f557d355241 sp:7ffc9fddc090 error:0 in 
    libdb-5.3.so[7f557d1fa000+1c2000]



The only segfault I see is:

  Apr 28 11:52:21 ... kernel: konsole[13357]: segfault at 60 ip 00007fecbd5ddc02
  sp 00007ffecee46cd8 error 4 in libQt5XcbQpa.so.5.9.4[7fecbd581000+f4000]

I have no idea whether that segfault would have occurred, too, if there hadn't been those problems with libdb-5.3.so

Closing as duplicate of bug 24767

*** This bug has been marked as a duplicate of bug 24767 ***

Resolution: (none) => DUPLICATE
Status: NEW => RESOLVED