Mageia forum

A few days ago the updater icon invitesd me to upgrade to kernel 5.5.4.
The two kernel RPMs were downloaded and installed, but after reboot the system still starts with the older kernel 5.4.17, as is conformed by the command: uname -a:

Code: Select all

$ uname -a
Linux kameel.artis 5.4.17-desktop-1.mga7 #1 SMP Sat Feb 1 21:57:04 UTC 2020 x86_64 x86_64 x86_64 GNU/Linux

I did a cold restart, but it stays in kernel 5.4.17.

I did urpmi --auto-update, to be sure the media are all up-to-date: Packages are up to date.

When starting the MCC, to check if kernel 5.5.4 is installed, 5.5.4 is indeed in the list of installed RPMs (green icon).
I removed the 5.5.4 RPM, rebooted, then installed 5.5.4 again. After restart, the old kernel 5.4.17 is still active after booting.

I'm puzzled what is wrong here? For many years I never had problems with updating to a new kernel.

How to fix this?

I'm happy to give more details from the logs, but don"t know where to look?

Marc.

Hi

can you report output of, as root:

Code: Select all

# update-grub2

.

Hi, here's the output. Thanks for looking at it.

I don't know what the error line "/dev/sdc: open failed: No medium found" means. The systems on sdc are working fine.
What means "DM multipath kernel driver not loaded"?

BTW, the last two lines are wrong: Mint is actually installed on /dev/sdc2, and Fedora is on /dev/sdc4.

Furthermore: my daily production system (another Mageia 7) on /dev/sdc1 is not found.

Code: Select all

# update-grub2
Generating grub configuration file ...
Found theme: /boot/grub2/themes/maggy/theme.txt
Found linux image: /boot/vmlinuz-5.5.4-desktop-1.mga7
Found initrd image: /boot/initrd-5.5.4-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.17-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.17-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.12-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.12-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.10-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.10-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.6-desktop-2.mga7
Found initrd image: /boot/initrd-5.4.6-desktop-2.mga7.img
Found linux image: /boot/vmlinuz-5.4.2-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.2-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.3.13-desktop-2.mga7
Found initrd image: /boot/initrd-5.3.13-desktop-2.mga7.img
Found linux image: /boot/vmlinuz-5.1.14-desktop-1.mga7
Found initrd image: /boot/initrd-5.1.14-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-desktop
Found initrd image: /boot/initrd-desktop.img
Feb 24 11:26:35 | DM multipath kernel driver not loaded
  /dev/sdc: open failed: No medium found
Found Mageia 7 (7) on /dev/sda3
Found Linux Mint 19.3 Tricia (19.3) on /dev/sdb2
Found Fedora 31 (Workstation Edition) on /dev/sdb4
done

I had the same problem today, couple of hours ego.
The output you're asking:

Code: Select all

# update-grub2
Generating grub configuration file ...
Found theme: /boot/grub2/themes/maggy/theme.txt
Found linux image: /boot/vmlinuz-5.4.17-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.17-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.12-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.12-desktop-1.mga7.img
Found linux image: /boot/vmlinuz-5.4.10-desktop-1.mga7
Found initrd image: /boot/initrd-5.4.10-desktop-1.mga7.img
Feb 24 12:23:17 | DM multipath kernel driver not loaded
Found Linux Mint 19.3 Tricia (19.3) on /dev/sdb1
done

(but on first reboot after update there was no option for the Mint installation in /dev/sdb1)

These are the packages that were in the update:
- calibre-3.42.0-3.1.mga7.x86_64
- cpupower-5.5.4-1.mga7.x86_64
- iproute2-5.5.0-1.mga7.x86_64
- kernel-desktop-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-devel-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-devel-latest-5.5.4-1.mga7.x86_64
- kernel-desktop-latest-5.5.4-1.mga7.x86_64
- kernel-userspace-headers-5.5.4-1.mga7.x86_64
- lib64systemd0-241-8.5.mga7.x86_64
- lib64udev1-241-8.5.mga7.x86_64
- nss-myhostname-241-8.5.mga7.x86_64
- patch-2.7.6-4.1.mga7.x86_64
- strace-5.5-1.mga7.x86_64
- systemd-241-8.5.mga7.x86_64
- systemd-units-241-8.5.mga7.x86_64
- virtualbox-kernel-5.5.4-desktop-1.mga7-6.0.16-5.mga7.x86_64
- virtualbox-kernel-desktop-latest-6.0.16-5.mga7.x86_64

But before that, i was asked to give my permission to remove all of nvidia proprietary drivers, modules, etc (sorry, I didn't copy that message).
Clicked OK, assuming that the next update would be the nvidia drivers.
The update finished, rebooted as prompted, the boot-up process stuck in "search for new hardware" a bit too long (4-5 minutes at least)
and when i got into KDE, there were no nvidia proprietary drivers of course, nouveau driver was used.
Launched MCC, manual update request, nothing proposed as update, then installed manually nvidia proprietary drivers (MCC>Hardware>Set up the graphical server),
rebooted as prompted, reboot stuck with the following message:

Code: Select all

[442.846974] nv_tco: Unexpected close, not stopping Watchdog!


So, I pressed the power button to shut-off the pc, booted again, show the message

Code: Select all

A start job is running for run dkms_autoinstaller on every …uild dkms modules for newly booted kernels


(that is not the message I see when nvidia proprietary drivers are compiled in the kernel)
When logged into KDE, still the old kernel is used.

Code: Select all

$ uname -r
5.4.17-desktop-1.mga7


Launched MCC, went to software Management, the following (kernel) packages are installed:
- kernel-desktop-devel-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-devel-latest-5.5.4-1.mga7.x86_64
but these are not installed:
- kernel-desktop-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-latest-5.5.4-1.mga7.x86_64
The strange thing is that these packages:
- kernel-desktop-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-latest-5.5.4-1.mga7.x86_64
are not listed at all (not even as uninstalled so I can check them for installation),
and no "kernel-desktop-latest" package is installed (the only packages present for installation
are for kernels 5.1.14 and 5.4.12).

Sorry for the long post, I tried to give as more information I could (and sorry for my english).

Something odd (or bad?) is happening with the Mageia 7 updates (that is not the first incident).

What to do now?
Is there a fix or workaround for today's problem as described from the op?

EDIT:
I also noticed these two packages are not installed:
- virtualbox-kernel-5.5.4-desktop-1.mga7-6.0.16-5.mga7.x86_64
- virtualbox-kernel-desktop-latest-6.0.16-5.mga7.x86_64
(they are not even listed as available), but the old ones are listed:
virtualbox-kernel-5.4.12-desktop-1.mga7
virtualbox-kernel-5.4.17-desktop-1.mga7
Also there are not any "virtualbox-kernel-desktop-latest" packages installed.

after latest update for me that includes kernel:

Code: Select all

$ uname -r
5.5.4-desktop-1.mga7

you both are multi booting, so which system is in control of the grub?
please boot into that one and update grub or grub2

marchugo wrote:Mint is actually installed on /dev/sdc2, and Fedora is on /dev/sdc4.

Furthermore: my daily production system (another Mageia 7) on /dev/sdc1 is not found.

2 possibilities:
motherboard has decided that your sdb is now sdc.
you may have a defective connection to one of your HDD.

(edit: What means "DM multipath kernel driver not loaded"?
this is an extraneous message, it can be safely ignored. end edit)

John66 wrote:A start job is running for run dkms_autoinstaller on every …uild dkms modules for newly booted kernels

this runs every boot to check if a new module is present but not installed.
if a new module is found, the driver is built and installed.
the nvidia update may not have been available when you updated and as your existing driver was not compatible with your new kernel, it was not installed.
(edit #2: this seems most likely and when I did my update, the virtualbox kernels were updated for me. end edit #2)

Thanks benmc for your reply.
Mageia 7 (in /dev/sda) controls the grub.
The output of "update-grub" posted in previous post.
The system seems to work OK, but that's not right.
The packages:
- kernel-desktop-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-latest-5.5.4-1.mga7.x86_64
are not even listed in software management.

The contents of /boot directory are:
dracut/
grub2/
config-5.4.10-desktop-1.mga7
config-5.4.12-desktop-1.mga7
config-5.4.17-desktop-1.mga7
initrd-5.4.10-desktop-1.mga7.img
initrd-5.4.12-desktop-1.mga7.img
initrd-5.4.17-desktop-1.mga7.img
symvers-5.4.10-desktop-1.mga7.xz
symvers-5.4.12-desktop-1.mga7.xz
symvers-5.4.17-desktop-1.mga7.xz
System.map-5.4.10-desktop-1.mga7
System.map-5.4.12-desktop-1.mga7
System.map-5.4.17-desktop-1.mga7
vmlinuz-5.4.10-desktop-1.mga7
vmlinuz-5.4.12-desktop-1.mga7
vmlinuz-5.4.17-desktop-1.mga7
There is no 5.5.4 kernel, how can this be? It was installed.

I'm thinking to unistall (tomorrow) the other 5.5.4 packages that were installed:
- cpupower-5.5.4-1.mga7.x86_64
- kernel-userspace-headers-5.5.4-1.mga7.x86_64
and see what happens.

Well, I fixed it.
First I uninstalled the package: kernel-userspace-headers-5.5.4-1.mga7.x86_64 (not sure if it made any diference at all).
Didn't uninstall the package: cpupower-5.5.4-1.mga7.x86_64 (too many dependencies to uninstall, I believe system would be broken).
Went to the update mirror I was using and in the updates section I found that there were not any updates about the 5.5.4 kernel.
So, the mirror was not updated.
Removed media sources, add sources from another mirror in another country (that the 5.5.4 kernel was present in core/updates),
did a manual update and gave me these updates:
- dkms-nvidia390-390.132-2.mga7.nonfree.x86_64
- lib64gd3-2.2.5-5.1.mga7.x86_64
- lib64xml2-devel-2.9.9-2.3.mga7.x86_64
- lib64xml2_2-2.9.9-2.3.mga7.x86_64
- libgbm1-19.3.4-1.mga7.i586
- libglapi0-19.3.4-1.mga7.i586
- libmesaegl1-19.3.4-1.mga7.i586
- libnspr4-4.25-1.mga7.i586
- libnss3-3.50.0-1.mga7.i586
- libqt5core5-5.12.6-2.mga7.i586
- libqt5dbus5-5.12.6-2.mga7.i586
- libqt5network5-5.12.6-2.mga7.i586
- libsystemd0-241-8.5.mga7.i586
- libudev1-241-8.5.mga7.i586
- libxml2-utils-2.9.9-2.3.mga7.x86_64
- libxml2_2-2.9.9-2.3.mga7.i586
- nvidia390-cuda-opencl-390.132-2.mga7.nonfree.x86_64
- nvidia390-doc-html-390.132-2.mga7.nonfree.x86_64
- radeon-firmware-20200204-1.mga7.nonfree.noarch
- wine32-4.0.3-1.mga7.i586
- x11-driver-video-nvidia390-390.132-2.mga7.nonfree.x86_64

Still, no updates for the kernel, but did the update anyway.
I manually downloaded the kernel update packages:
- kernel-desktop-5.5.4-1.mga7-1-1.mga7.x86_64.rpm
- kernel-desktop-latest-5.5.4-1.mga7.x86_64.rpm
- kernel-desktop-devel-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-devel-latest-5.5.4-1.mga7.x86_64
Tried to install them manually in terminal, the last two wouldn't install (they conflicted each other)
I installed the first two, rebooted, reboot stuck again, with the message

Code: Select all

nv_tco: Unexpected close, not stopping Watchdog!

Cold-boot (pc power button), when logged in KDE the 5.5.4 kernel was running, but the nouveau driver was in use not the proprietary nvidia one.
Went to MCC>Software>Software Management and noticed that the packages:
- kernel-desktop-devel-5.5.4-1.mga7-1-1.mga7.x86_64
- kernel-desktop-devel-latest-5.5.4-1.mga7.x86_64
were unchecked. So I checked them for installation, then I rebooted the pc, again stuck with the message:

Code: Select all

nv_tco: Unexpected close, not stopping Watchdog!

During boot process I noticed the message "Installing nvidia 390 drivers", but when logged in, still the nouveau drivers were in use.
I launched MCC>Software>Software Management but the nvidia proprietary drivers were listed as installed.
So, I went to MCC>Hardware>Set up the graphical server and set the nvidia proprietary drivers to be used, rebooted (no error message this time)
and all went fine.
Installed also the packages:
- virtualbox-kernel-5.5.4-desktop-1.mga7-6.0.16-5.mga7.x86_64
- virtualbox-kernel-desktop-latest-6.0.16-5.mga7.x86_64
that were not listed as installed (although they were initially) and everything is as it should be.

So, it seems that (maybe) the problem started with the not updated mirror (but how the updates even came and installed if they were not present
in the core/update media, I can't understand it).

Anyway, everything is good now.

John66 ,
thats great.
hope everything is now in order.

marchugo:

I am concerned about your missing "other" Mageia install
can you post ouput of :

Code: Select all

lsblk

if your "other" Mageia was in control of your boot, then grub would not have been updated, even though a new config file was generated with both the new kernel and by manually running "update-grub2"

the following is assuming BIOS boot, not EFI:
to make your current Mageia be "in control" of grub, you will need to run the following as root

Code: Select all

# grub2-install /dev/sdx

where "x" is your boot device as set in your bios.

if your system is set up for EFI, I will have to dig up the correct command

Hi benmc,

Thanks for your time.

Here's the output of lsblk:

Code: Select all

$ lsblk
NAME   MAJ:MIN RM   SIZE RO TYPE MOUNTPOINT
sda      8:0    0 189,9G  0 disk
├─sda1   8:1    0  29,3G  0 part
├─sda2   8:2    0   251M  0 part
├─sda3   8:3    0 159,4G  0 part
└─sda4   8:4    0 996,2M  0 part [SWAP]
sdb      8:16   0 465,8G  0 disk
├─sdb1   8:17   0 270,1G  0 part /
├─sdb2   8:18   0  37,3G  0 part
├─sdb3   8:19   0     1K  0 part
├─sdb4   8:20   0  63,2G  0 part
└─sdb5   8:21   0  37,3G  0 part
sr0     11:0    1  1024M  0 rom 

For some strange reason during the install of MGA 7.1 on my SSD drive, this drive got named /dev/sdc. In the boot menu at cold restart al the options (MGA, Mint, Fedora) still have the name /sdc in them. So there must be some grub file somewhere that is called during restart. I don't know where it is.

For another unknown reason this SSD drive is now called /dev/sdb. All the systems on that drive can start normally.

I stumbled upon another problem today, using the command fdisk -l: the partition /sdb3 is a mess, it doesn't start at a correct physical sector boundary (see output). In fact, it doesn't start at all anymore. sdb3 contains a second install of MGA7, that I was trying. Maybe that is the heart of the problem with updating the kernel? Don't know how this misalignment happened, I created all the partitions with the latest version of GParted, before installing the systems on them. I'm not sure if this misalignment plays any role in the problem with the not updating kernel 5.5.4. As a layman I'm not confident enough to repair this.

Code: Select all

$ fdisk -l
Disk /dev/sda: 189,9 GiB, 203928109056 bytes, 398297088 sectors
Disk model: Maxtor 6L200M0 
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 512 bytes
I/O size (minimum/optimal): 512 bytes / 512 bytes
Disklabel type: dos
Disk identifier: 0x0009b4f1

Device     Boot    Start       End   Sectors   Size Id Type
/dev/sda1  *          63  61432559  61432497  29,3G  b W95 FAT32
/dev/sda2       61432560  61946639    514080   251M 83 Linux
/dev/sda3       63986895 398283479 334296585 159,4G 83 Linux
/dev/sda4       61946640  63986894   2040255 996,2M 82 Linux swap / Solaris

Partition table entries are not in disk order.


Disk /dev/sdb: 465,8 GiB, 500107862016 bytes, 976773168 sectors
Disk model: CT500MX500SSD1 
Units: sectors of 1 * 512 = 512 bytes
Sector size (logical/physical): 512 bytes / 4096 bytes
I/O size (minimum/optimal): 4096 bytes / 4096 bytes
Disklabel type: dos
Disk identifier: 0x553205cd

Device     Boot     Start       End   Sectors   Size Id Type
/dev/sdb1  *         2048 566530047 566528000 270,1G 83 Linux
/dev/sdb2       566530048 644655047  78125000  37,3G 83 Linux
/dev/sdb3       644655102 722780159  78125058  37,3G  5 Extended
/dev/sdb4       722780160 855320167 132540008  63,2G 83 Linux
/dev/sdb5       644655104 722780159  78125056  37,3G 83 Linux

Partition 3 does not start on physical sector boundary.
Partition table entries are not in disk order.


marchugo wrote: menu at cold restart al the options (MGA, Mint, Fedora) still have the name /sdc in them

this is consistent with grub menu not being updated.
boot into and update grub in that other Mga.
that should fix your boot entries.

otherwise, run that grub2-install command.

it would appear that at some time past you had a 3rd HDD: sda, it since been removed. and sdb, sdc remain.

re:

marchugo wrote:Partition 3 does not start on physical sector boundary

"if it aint broke, dont fix it."
apart from that, I have found it does occur a lot with Gparted.
With a lot of work, you can adjust the sizes of your partitions to align the partition with the sector boundary.

I have found that diskdrake is much better at sector boundary alignments when partitioning,however, as always YMMV

@benmc
The command grub2-install /dev/sdb did work. Now the bootmenu is OK. The weird thing is that the Welcome page now shows that kernel 5.5.4 is indeed already installed.

The downside of using the grub2-install command is, that Fedora on partition /dev/sdb4 is now corrupt. Booting it stops halfway with a lot of errors.
I now have to decide if I re-install Fedora again, at the risk that the grub will be ruined again.
MGA 7 and Mint are working fine.

Thanks a lot for helping me out.
I will report what happens if I reinstall Fedora.

Marc.