On 14 May my usual "update everything" from the Princeton math server updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to choose
its server from the mirror list, and that has provided an update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc on
my backup machine.
On 2021-05-23, Jim <jim.beard@verizon.net> wrote:
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to choose
its server from the mirror list, and that has provided an update to
kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc on
my backup machine.
It is in backports. You must have backports as part of your list of repositories in your main machine as well as release and updates
.
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to choose
its server from the mirror list, and that has provided an update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc on
my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8
cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the final -1-1.mga8.
On Sun, 23 May 2021 12:32:30 -0400, Jim <jim.beard@verizon.net> wrote:backports/kernel-desktop-5.12.4-1.mga8-1-1.mga8.x86_64.rpm
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to choose
its server from the mirror list, and that has provided an update to
kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc
on my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields
kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8 cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the final
-1-1.mga8.
http://mirror.math.princeton.edu/pub/mageia/distrib/8/x86_64/media/core/
The backports repos should only be enabled, an intentionally chosen
package installed and then the backports repos should be disabled.
Backports should not be use as a source for auto-select.
On Sun, 23 May 2021 15:01:11 -0400, David W. Hodgins wrote:core/
On Sun, 23 May 2021 12:32:30 -0400, Jim <jim.beard@verizon.net> wrote:
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to
choose its server from the mirror list, and that has provided an
update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc
on my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields
kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8 cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of
kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the
final -1-1.mga8.
http://mirror.math.princeton.edu/pub/mageia/distrib/8/x86_64/media/
backports/kernel-desktop-5.12.4-1.mga8-1-1.mga8.x86_64.rpm
The backports repos should only be enabled, an intentionally chosen
package installed and then the backports repos should be disabled.
Backports should not be use as a source for auto-select.
I will accept the recommendation from you and William Unruh. Backports disabled.
Years ago, when applications were evolving rapidly, I got in the habit
of updating with any backports available, and it served me will, in my estimation. These days, it would seem unneeded.
FWIW, Pan on my main machine running 5.12.4-1 is not able to contact eternal-september.org, but this machine my backup just updated to the 5.10.37-2 kernel is working fine.
I tinkered with pan, changing the backup server when pan complained it
could not connect to reader80.eternal-september.org on port 80
(plaintext) to reader443.eternal-september.org on port 443 (encrypted), Perhaps that backup should be removed entirely, as port 563 (encryted)
alone works fine with this machine.
Or perhaps the backport 5.12.4-1 kernel has a kink or two. Deleting it
and updating to 5.10.37-2 would be a possibility....
Cheers!
jim b.
On Sun, 23 May 2021 15:01:11 -0400, David W. Hodgins wrote:core/
On Sun, 23 May 2021 12:32:30 -0400, Jim <jim.beard@verizon.net> wrote:
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to
choose its server from the mirror list, and that has provided an
update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc
on my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields
kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8 cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of
kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the
final -1-1.mga8.
http://mirror.math.princeton.edu/pub/mageia/distrib/8/x86_64/media/
backports/kernel-desktop-5.12.4-1.mga8-1-1.mga8.x86_64.rpm
The backports repos should only be enabled, an intentionally chosen
package installed and then the backports repos should be disabled.
Backports should not be use as a source for auto-select.
I will accept the recommendation from you and William Unruh. Backports disabled.
Years ago, when applications were evolving rapidly, I got in the habit
of updating with any backports available, and it served me will, in my estimation. These days, it would seem unneeded.
FWIW, Pan on my main machine running 5.12.4-1 is not able to contact eternal-september.org, but this machine my backup just updated to the 5.10.37-2 kernel is working fine.
I tinkered with pan, changing the backup server when pan complained it
could not connect to reader80.eternal-september.org on port 80
(plaintext) to reader443.eternal-september.org on port 443 (encrypted), Perhaps that backup should be removed entirely, as port 563 (encryted)
alone works fine with this machine.
Or perhaps the backport 5.12.4-1 kernel has a kink or two. Deleting it
and updating to 5.10.37-2 would be a possibility....
On Mon, 24 May 2021 00:34:04 +0000, Jim Beard wrote:
On Sun, 23 May 2021 15:01:11 -0400, David W. Hodgins wrote:core/
On Sun, 23 May 2021 12:32:30 -0400, Jim <jim.beard@verizon.net> wrote:
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to
choose its server from the mirror list, and that has provided an
update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc
on my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields
kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8 cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of
kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the
final -1-1.mga8.
http://mirror.math.princeton.edu/pub/mageia/distrib/8/x86_64/media/
backports/kernel-desktop-5.12.4-1.mga8-1-1.mga8.x86_64.rpm
The backports repos should only be enabled, an intentionally chosen
package installed and then the backports repos should be disabled.
Backports should not be use as a source for auto-select.
I will accept the recommendation from you and William Unruh. Backports
disabled.
Years ago, when applications were evolving rapidly, I got in the habit
of updating with any backports available, and it served me will, in my
estimation. These days, it would seem unneeded.
FWIW, Pan on my main machine running 5.12.4-1 is not able to contact
eternal-september.org, but this machine my backup just updated to the
5.10.37-2 kernel is working fine.
I tinkered with pan, changing the backup server when pan complained it
could not connect to reader80.eternal-september.org on port 80
(plaintext) to reader443.eternal-september.org on port 443 (encrypted),
Perhaps that backup should be removed entirely, as port 563 (encryted)
alone works fine with this machine.
Or perhaps the backport 5.12.4-1 kernel has a kink or two. Deleting it
and updating to 5.10.37-2 would be a possibility....
continuing the saga,
I rebooted my main machine to 5.10.33 kernel, used urpme to remove the
5.12.4 kernel (which took another three rpms with it), and tried to
update, expecting the 5.10.37-2 kernel to be installed. It was not.
I went in mcc, deleted anything 5.12.4 that I thought non-critical
and then chose the 5.10.37-2 kernel to install. That brought in well
over a hundred rpms as dependencies, and also demanded to install the
5.12.4 kernel as a dependency.
At this point, I decided to be adventurous, and went for it.
That accomplished I rebooted to 5.10.37-2 and went into mcc and tried to delete the 5.12.4 kernel. This time it deleted the thing.
There is still a cpu-5.12.4 rpm that I had not deleted but that no longer shows up among my system's rpms, nor is it available via mcc.
Fingers crossed, I shall continue onward, hoping all will work.
Moral: A backport kernel may involve more complexities than one may wish
to deal with.
Cheers!
jim b.
On 5/23/21 9:40 PM, Jim wrote:
On Mon, 24 May 2021 00:34:04 +0000, Jim Beard wrote:If you delete the highest numbered kernel and its kernel-devel
On Sun, 23 May 2021 15:01:11 -0400, David W. Hodgins wrote:core/
On Sun, 23 May 2021 12:32:30 -0400, Jim <jim.beard@verizon.net>
wrote:
On 14 May my usual "update everything" from the Princeton math
server updated my main machine to
kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to
choose its server from the mirror list, and that has provided an
update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by
mcc on my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields
kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8 cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of
kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the
final -1-1.mga8.
http://mirror.math.princeton.edu/pub/mageia/distrib/8/x86_64/media/
backports/kernel-desktop-5.12.4-1.mga8-1-1.mga8.x86_64.rpm
The backports repos should only be enabled, an intentionally chosen
package installed and then the backports repos should be disabled.
Backports should not be use as a source for auto-select.
I will accept the recommendation from you and William Unruh.
Backports disabled.
Years ago, when applications were evolving rapidly, I got in the habit
of updating with any backports available, and it served me will, in my
estimation. These days, it would seem unneeded.
FWIW, Pan on my main machine running 5.12.4-1 is not able to contact
eternal-september.org, but this machine my backup just updated to the
5.10.37-2 kernel is working fine.
I tinkered with pan, changing the backup server when pan complained it
could not connect to reader80.eternal-september.org on port 80
(plaintext) to reader443.eternal-september.org on port 443
(encrypted),
Perhaps that backup should be removed entirely, as port 563 (encryted)
alone works fine with this machine.
Or perhaps the backport 5.12.4-1 kernel has a kink or two. Deleting
it and updating to 5.10.37-2 would be a possibility....
continuing the saga,
I rebooted my main machine to 5.10.33 kernel, used urpme to remove the
5.12.4 kernel (which took another three rpms with it), and tried to
update, expecting the 5.10.37-2 kernel to be installed. It was not.
I went in mcc, deleted anything 5.12.4 that I thought non-critical and
then chose the 5.10.37-2 kernel to install. That brought in well over
a hundred rpms as dependencies, and also demanded to install the 5.12.4
kernel as a dependency.
At this point, I decided to be adventurous, and went for it.
That accomplished I rebooted to 5.10.37-2 and went into mcc and tried
to delete the 5.12.4 kernel. This time it deleted the thing.
There is still a cpu-5.12.4 rpm that I had not deleted but that no
longer shows up among my system's rpms, nor is it available via mcc.
Fingers crossed, I shall continue onward, hoping all will work.
Moral: A backport kernel may involve more complexities than one may
wish to deal with.
Cheers!
jim b.
packages(if used), then you need to install the kernel-whatever-latest
and kernel-whatever-devel-latest for the now highest numbered
kernel(whatever being desktop, server, or desktop586, whichever is
installed) before that kernel will update. Also, if you are using
virtualbox, you'll need the proper "latest" kernel modules for that,
too.
The "latest" packages are what ensures you get the newest kernel updates
when they become available. They are replaced with new ones when you
update through normal update means, but not if you just install that
kernel on your own.
BTW, I haven't looked today, but yesterday there was a new kernel 5.12,
I think 5.12.6, in backports-testing, awaiting QA attention...
I think I will leave backports-testing for those interested in testing.
Even when I installed backports, I did not go for testing.
On Wed, 26 May 2021 14:11:07 -0000 (UTC), Jim wrote:Of course you could just put "ignore" into the stanze for those options
I think I will leave backports-testing for those interested in testing.
Even when I installed backports, I did not go for testing.
Once I pick/set a mirror I will run
urpmi.removemedia -y Debug Backport Testing 32bit
to remove those selections and all 32bit media since I do not need any 32 apps.
mcc used to install kernel-desktop-latest-5.10.37-2.mga8, which was
missing.
kernel-desktop-devel-latest-5.10.37-2.mga8 was installed when I
installed kernel-desktop-5.10.37-2.mga8-1-1.mga8.
I think I will leave backports-testing for those interested in testing.
Even when I installed backports, I did not go for testing.
On 5/26/21 10:11 AM, Jim wrote:
mcc used to install kernel-desktop-latest-5.10.37-2.mga8, which was
missing.
kernel-desktop-devel-latest-5.10.37-2.mga8 was installed when I
installed kernel-desktop-5.10.37-2.mga8-1-1.mga8.
I think I will leave backports-testing for those interested in testing.
Even when I installed backports, I did not go for testing.
The kernel-xx is a dependency of kernel-latest-xx, but kernel-latest-xx
is not a dependency of kernel-xx, so it is possible to have a kernel installed without the associated "latest" package.
When you installed/upgraded to Mageia 8, the kernel-latest associated
with that kernel was installed automatically. When you update a kernel through MCC, the old kernel-latest is replaced with the newer one, but
the old kernel is not removed, so that you have a backup in case the new kernel doesn't work for you.
When you removed the backported kernel, the associated kernel-latest was
also removed, because it depends on the kernel. But, the kernel-latest
that was removed when the backported kernel was installed is *NOT* automatically re-installed. You have to do that manually. If you don't,
then any new kernels that come along will not be pulled in by MCC as
updates.
On Fri, 28 May 2021 08:22:39 -0400, TJ wrote:
On 5/26/21 10:11 AM, Jim wrote:
mcc used to install kernel-desktop-latest-5.10.37-2.mga8, which was
missing.
kernel-desktop-devel-latest-5.10.37-2.mga8 was installed when I
installed kernel-desktop-5.10.37-2.mga8-1-1.mga8.
I think I will leave backports-testing for those interested in testing.
Even when I installed backports, I did not go for testing.
The kernel-xx is a dependency of kernel-latest-xx, but kernel-latest-xx
is not a dependency of kernel-xx, so it is possible to have a kernel
installed without the associated "latest" package.
When you installed/upgraded to Mageia 8, the kernel-latest associated
with that kernel was installed automatically. When you update a kernel
through MCC, the old kernel-latest is replaced with the newer one, but
the old kernel is not removed, so that you have a backup in case the new
kernel doesn't work for you.
When you removed the backported kernel, the associated kernel-latest was
also removed, because it depends on the kernel. But, the kernel-latest
that was removed when the backported kernel was installed is *NOT*
automatically re-installed. You have to do that manually. If you don't,
then any new kernels that come along will not be pulled in by MCC as
updates.
Thank you.
I now have,
rpm -qa |grep latest
kernel-desktop-devel-latest-5.10.37-2.mga8 kernel-desktop-latest-5.10.37-2.mga8
I think this should do it.
Cheers!
jim .
On 14 May my usual "update everything" from the Princeton math server
updated my main machine to kernel-desktop-5.12.4-1.mga8-1-1.mga8.
Today I got around to updating my backup machine, allowing it to choose
its server from the mirror list, and that has provided an update to kernel-desktop-5.10.37.4-2.mga8-1-1.mga8.
I used mcc to look at the princeton server, and 5.12.4-1 is there,
but did not appear when I searched the "mirror media" provided by mcc on
my backup machine.
An oddity I had not noticed, running rpm -qa |grep 5.12.4 yields kernel-userspace-headers-5.12.4-1.mga8
kernel-desktop-5.12.4-1.mga8-1-1.mga8
cpupower-5.12.4-1.mga8
kernel-desktop-latest-5.12.4-1.mga8
Note the -1-1.mga8 at the end of kernel-desktop-5.12.4-1.mga8-1-1.mga8.
A search using mcc shows kernel-desktop-5.12.4-1.mga8 without the final -1-1.mga8.
Does anyone know what is going on with this?
Cheers!
jim b.
On 5/28/21 10:36 PM, Jim wrote:
On Fri, 28 May 2021 08:22:39 -0400, TJ wrote:Just one more thing...
On 5/26/21 10:11 AM, Jim wrote:
mcc used to install kernel-desktop-latest-5.10.37-2.mga8, which was
missing.
kernel-desktop-devel-latest-5.10.37-2.mga8 was installed when I
installed kernel-desktop-5.10.37-2.mga8-1-1.mga8.
I think I will leave backports-testing for those interested in
testing.
Even when I installed backports, I did not go for testing.
The kernel-xx is a dependency of kernel-latest-xx, but
kernel-latest-xx is not a dependency of kernel-xx, so it is possible
to have a kernel installed without the associated "latest" package.
When you installed/upgraded to Mageia 8, the kernel-latest associated
with that kernel was installed automatically. When you update a kernel
through MCC, the old kernel-latest is replaced with the newer one, but
the old kernel is not removed, so that you have a backup in case the
new kernel doesn't work for you.
When you removed the backported kernel, the associated kernel-latest
was also removed, because it depends on the kernel. But, the
kernel-latest that was removed when the backported kernel was
installed is *NOT* automatically re-installed. You have to do that
manually. If you don't, then any new kernels that come along will not
be pulled in by MCC as updates.
Thank you.
I now have,
rpm -qa |grep latest kernel-desktop-devel-latest-5.10.37-2.mga8
kernel-desktop-latest-5.10.37-2.mga8
I think this should do it.
Cheers!
jim .
Every kernel update also includes an update to the cpupower package,
which, as I misunderstand it, is a collection of tools for tweaking
power management settings. Now, I don't know if you use that or not, (I don't, at least not consciously) but logic would lead me to believe that
the cpupower for the kernel 5.12 series could easily have issues here
and there if you attempt to use it with a 5.10 series kernel.
In addition, when the next 5.10 series update, now in testing, comes
along, if the cpupower for 5.12 is in place, it will not be updated by
MCC, because it has a higher number than the proposed update.
So just to be safe, you should probably downgrade cpupower so it matches
your current kernel-latest, as well.
| Sysop: | Keyop |
|---|---|
| Location: | Huddersfield, West Yorkshire, UK |
| Users: | 475 |
| Nodes: | 16 (2 / 14) |
| Uptime: | 141:45:47 |
| Calls: | 9,474 |
| Calls today: | 5 |
| Files: | 13,610 |
| Messages: | 6,120,116 |