I used the CLI update process, and so far things are looking good. But
when my rEFInd came up I noticed 'Mga7'. When it booted it's definitely
Mga8, but it's still using a Mga7 kernel. The Mga8 kernel is installed:
how do I force it to use the new kernel?
On Thu, 11 Mar 2021 13:08:12 -0500, chris <invalid@invalid.com> wrote:
I used the CLI update process, and so far things are looking good. But
when my rEFInd came up I noticed 'Mga7'. When it booted it's definitely
Mga8, but it's still using a Mga7 kernel. The Mga8 kernel is installed:
how do I force it to use the new kernel?
At the refind boot selection screen press any of the insert, tab, or f2
keys
for more options. Select the linux entry rather then a specific vmlinuz key so it will use the latest installed kernel. Once all hardware confirmed
to be
working ok, uninstall the mga7 kernel packages.
Regards, Dave Hodgins
On 11/03/2021 20:18, David W. Hodgins wrote:
On Thu, 11 Mar 2021 13:08:12 -0500, chris <invalid@invalid.com> wrote:
I used the CLI update process, and so far things are looking good. But
when my rEFInd came up I noticed 'Mga7'. When it booted it's definitely
Mga8, but it's still using a Mga7 kernel. The Mga8 kernel is installed:
how do I force it to use the new kernel?
At the refind boot selection screen press any of the insert, tab, or f2
keys
for more options. Select the linux entry rather then a specific vmlinuz key >> so it will use the latest installed kernel. Once all hardware confirmed
to be
working ok, uninstall the mga7 kernel packages.
Regards, Dave Hodgins
Thanks, Dave. Will give that a go.
I've been digging through the refind pages to try to understande better.
One thing surprised me: when I looked at refind_linux.conf it gave:
[root@localhost boot]# cat refind_linux.conf
"linux" "root=/dev/nvme0n1p7 splash quiet noiswmd resume=/dev/nvme0n1p6 audit=0 vga=791"
"linux-nonfb" "root=/dev/nvme0n1p7 resume=/dev/nvme0n1p6"
"failsafe" "root=/dev/nvme0n1p7 failsafe noiswmd audit=0"
Those 'resume' options are pointing to my swap partition. That seems odd
to me!
On 2021-03-11, chris <invalid@invalid.com> wrote:Ah thanks. That makes sense. I just wanted to check refind hadn't done something peculiar!
On 11/03/2021 20:18, David W. Hodgins wrote:
On Thu, 11 Mar 2021 13:08:12 -0500, chris <invalid@invalid.com> wrote:
I used the CLI update process, and so far things are looking good. But >>>> when my rEFInd came up I noticed 'Mga7'. When it booted it's definitely >>>> Mga8, but it's still using a Mga7 kernel. The Mga8 kernel is installed: >>>> how do I force it to use the new kernel?
At the refind boot selection screen press any of the insert, tab, or f2
keys
for more options. Select the linux entry rather then a specific vmlinuz key >>> so it will use the latest installed kernel. Once all hardware confirmed
to be
working ok, uninstall the mga7 kernel packages.
Regards, Dave Hodgins
Thanks, Dave. Will give that a go.
I've been digging through the refind pages to try to understande better.
One thing surprised me: when I looked at refind_linux.conf it gave:
[root@localhost boot]# cat refind_linux.conf
"linux" "root=/dev/nvme0n1p7 splash quiet noiswmd resume=/dev/nvme0n1p6
audit=0 vga=791"
"linux-nonfb" "root=/dev/nvme0n1p7 resume=/dev/nvme0n1p6"
"failsafe" "root=/dev/nvme0n1p7 failsafe noiswmd audit=0"
Those 'resume' options are pointing to my swap partition. That seems odd
to me!
hibernate and sleep images are stored in swap. This says that if you are rebooting into one of those, get the stuff from swap.
hibernate and sleep images are stored in swap. This says that if you are rebooting into one of those, get the stuff from swap.
On Thu, 11 Mar 2021 13:08:12 -0500, chris <invalid@invalid.com> wrote:
I used the CLI update process, and so far things are looking good. But
when my rEFInd came up I noticed 'Mga7'. When it booted it's definitely
Mga8, but it's still using a Mga7 kernel. The Mga8 kernel is installed:
how do I force it to use the new kernel?
At the refind boot selection screen press any of the insert, tab, or f2
keys
for more options. Select the linux entry rather then a specific vmlinuz key so it will use the latest installed kernel. Once all hardware confirmed
to be
working ok, uninstall the mga7 kernel packages.
Regards, Dave Hodgins
OK I'm now booting into the right kernel, but the refind boot menu still offers two boot icons. Going into them shows options that include Mga8,
but also trying to boot using initrd.old (which had pointed to mga7).
How do I edit the menu to get rid of surplus and unusable options? I've looked at refind.conf but that doesn't seem to include my specific menu
items - it only gives examples of how to do them. Where are they saved
so I can edit them, manually?
On Fri, 12 Mar 2021 08:18:18 -0500, chris <invalid@invalid.com> wrote:
OK I'm now booting into the right kernel, but the refind boot menu still
offers two boot icons. Going into them shows options that include Mga8,
but also trying to boot using initrd.old (which had pointed to mga7).
The Mageia control center will only update the boot configuration for the boot loader selected in mmc, Boot, Set up boot system, during a kernel install.
Either remember to run (as root) update-grub after every kernel install, or un-install the package grub2-efi.
The two icons shown by refind are for booting directly into the selected kernel,
or to boot using grub2.
How do I edit the menu to get rid of surplus and unusable options? I've
looked at refind.conf but that doesn't seem to include my specific menu
items - it only gives examples of how to do them. Where are they saved
so I can edit them, manually?
The refind kernel boot parameters are stored in /boot/refind_linux.conf.
On 12/03/2021 14:36, David W. Hodgins wrote:
Thanks for this: I realised I had a misunderstanding that grub had been replaced by EFI go was avoiding grub completely!
The two icons shown by refind are for booting directly into the selected
kernel,
or to boot using grub2.
The grub option on refind is now as it should be and boots. However...
How do I edit the menu to get rid of surplus and unusable options? I've
looked at refind.conf but that doesn't seem to include my specific menu
items - it only gives examples of how to do them. Where are they saved
so I can edit them, manually?
The refind kernel boot parameters are stored in /boot/refind_linux.conf.
The refind option has 'Boot boot\vmlinuz.old' - this can't be right?
1.8969 systemd[1]: Failed to look up module alias 'autofs4'... Failed to start Load legacy module configuration
I suspect it might be easiest to remove refind entirely (and maybe
reinstall) but I wanted to try and resolve these issues (not finding any relevant help online).
On Fri, 12 Mar 2021 15:59:18 -0500, chris <invalid@invalid.com> wrote:
On 12/03/2021 14:36, David W. Hodgins wrote:
Thanks for this: I realised I had a misunderstanding that grub had been
replaced by EFI go was avoiding grub completely!
The two icons shown by refind are for booting directly into the selected >>> kernel,
or to boot using grub2.
The grub option on refind is now as it should be and boots. However...
How do I edit the menu to get rid of surplus and unusable options? I've >>>> looked at refind.conf but that doesn't seem to include my specific menu >>>> items - it only gives examples of how to do them. Where are they saved >>>> so I can edit them, manually?
The refind kernel boot parameters are stored in /boot/refind_linux.conf.
The refind option has 'Boot boot\vmlinuz.old' - this can't be right?
Ah. I thought you were referring to options passed on the command line
to the
kernel and/or systemd, but you're referring to the choices of kernels to boot.
With grub2-efi it uses a menu in /boot/grub2 to show the kernels
available for
selection and the contents of /etc/default/grub to control what kernel parameters
are passed. As you've found, if the menu doesn't match what's available
it causes
problems.
With refind, there is no file with a menu that has to be maintained. At
boot it
searches normal locations for a kernel (linux or windows) to boot and
builds the
selection menu dynamically. For linux boots, it uses refind_linux.conf
only to
control what kernel parameters are passed on the command line to the selected
kernel It considers each filename in /boot that starts with vmlinuz to
be a kernel
that can be selected.
The linux, linux-nonfb, and failsafe are different combinations of
kernel parameters
that will be passed to the kernel. Since refind_linux.conf has three
options listed
within it, each kernel has three options selectable at boot.
The choices for kernels is based on "ls -l /boot/vmlinuz*".
In Mageia, the first vmlinuz is a symlink to the latest kernel
installed, and that
is what should normally be used.
To clean up the list, uninstall any old kernels and remove any manually created
files that start with vmlinuz in /boot. If you want to keep vmlinuz.old, rename it to something like was.vmlinuz.old so refind won't see it as a kernel.
The old kernels can be uninstalled using mcc or using urpmi. I use a
script,
https://bugs.mageia.org/attachment.cgi?id=10860 to uninstall the old
kernels and
related packages.
1.8969 systemd[1]: Failed to look up module alias 'autofs4'... Failed to
start Load legacy module configuration
That's actually a normal error and doesn't matter how the system is booted. See https://bugs.mageia.org/show_bug.cgi?id=19795
I suspect it might be easiest to remove refind entirely (and maybe
reinstall) but I wanted to try and resolve these issues (not finding any
relevant help online).
I find it works very well, and is much bettor on systems with multiple
linux
installs. I'd try cleaning up the kernels and try it for a while before deciding.
Regards, Dave Hodgins
On 12/03/2021 22:30, David W. Hodgins wrote:
On Fri, 12 Mar 2021 15:59:18 -0500, chris <invalid@invalid.com> wrote:
On 12/03/2021 14:36, David W. Hodgins wrote:
Thanks for this: I realised I had a misunderstanding that grub had been
replaced by EFI go was avoiding grub completely!
The two icons shown by refind are for booting directly into the
selected
kernel,
or to boot using grub2.
The grub option on refind is now as it should be and boots. However...
How do I edit the menu to get rid of surplus and unusable options?
I've
looked at refind.conf but that doesn't seem to include my specific
menu
items - it only gives examples of how to do them. Where are they saved >>>>> so I can edit them, manually?
The refind kernel boot parameters are stored in
/boot/refind_linux.conf.
The refind option has 'Boot boot\vmlinuz.old' - this can't be right?
Ah. I thought you were referring to options passed on the command line
to the
kernel and/or systemd, but you're referring to the choices of kernels
to boot.
With grub2-efi it uses a menu in /boot/grub2 to show the kernels
available for
selection and the contents of /etc/default/grub to control what kernel
parameters
are passed. As you've found, if the menu doesn't match what's
available it causes
problems.
With refind, there is no file with a menu that has to be maintained.
At boot it
searches normal locations for a kernel (linux or windows) to boot and
builds the
selection menu dynamically. For linux boots, it uses refind_linux.conf
only to
control what kernel parameters are passed on the command line to the
selected
kernel It considers each filename in /boot that starts with vmlinuz to
be a kernel
that can be selected.
The linux, linux-nonfb, and failsafe are different combinations of
kernel parameters
that will be passed to the kernel. Since refind_linux.conf has three
options listed
within it, each kernel has three options selectable at boot.
The choices for kernels is based on "ls -l /boot/vmlinuz*".
In Mageia, the first vmlinuz is a symlink to the latest kernel
installed, and that
is what should normally be used.
To clean up the list, uninstall any old kernels and remove any
manually created
files that start with vmlinuz in /boot. If you want to keep vmlinuz.old,
rename it to something like was.vmlinuz.old so refind won't see it as
a kernel.
The old kernels can be uninstalled using mcc or using urpmi. I use a
script,
https://bugs.mageia.org/attachment.cgi?id=10860 to uninstall the old
kernels and
related packages.
Thanks for that. I didn't want to rename vmlinuz.old in case it broke something else! But I now have only the one mga8 kernel installed.
Sorry, I should have added: those messages came just before saying it1.8969 systemd[1]: Failed to look up module alias 'autofs4'... Failed to >>> start Load legacy module configuration
That's actually a normal error and doesn't matter how the system is
booted.
See https://bugs.mageia.org/show_bug.cgi?id=19795
was dropping back to emergency mode. It didn't get any further, so a
reboot was necessary.
I currently only have Mga8 installed, so refind isn't strictlyI suspect it might be easiest to remove refind entirely (and maybe
reinstall) but I wanted to try and resolve these issues (not finding any >>> relevant help online).
I find it works very well, and is much bettor on systems with multiple
linux
installs. I'd try cleaning up the kernels and try it for a while
before deciding.
necessary, but I do want to get it working properly for my own
satisfaction.
Regards, Dave Hodgins
OK all now seems to be fine on the refind/boot front. Thanks for your patience.
Now all I have to do is work out why shutdown is delayed by
Failed unmounting /home
Failed unmounting Temporary Directory
and systemd-shutdown[1]: Waiting for process: java, wrapper-linux
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 371 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 172:23:49 |
| Calls: | 7,915 |
| Files: | 12,983 |
| Messages: | 5,797,535 |