Updating @world set is now optional?

[st834]

Hi,

I use the Gentoo Handbook (https://wiki.gentoo.org/wiki/Handbook:AMD64) as an instruction for installing Gentoo.

In the "Installing the Gentoo base system" step (https://wiki.gentoo.org/wiki/Handbook:A ... ation/Base), update the system

Code: Select all

emerge --ask --verbose --update --deep --newuse @world

after you select and install a profile in the new revision of Handbook, it is marked as optional. In the previous edition, it was mandatory. Unfortunately, I could not attach screenshots here, but this can be seen from the table of contents and the text of the page to which I cited the link.

I decided to formally follow this guide and missed the @world set update step and moved on to the next step - kernel setup.

It turned out that without updating the @world set, I do not have fundamental packages (for example, systemd), without which it is impossible to perform further steps.

Question: Why is the update of the @ world kit now marked as optional in the new version of the installation guide? Maybe I misunderstood the Handbook? If I were installing Gentoo for the first time, I would no doubt miss this step, as it is now described.

I am grateful for any clarification!

[logrusx]

You most likely chose a stage3 tarball that didn't match the profile you later selected.

Best Regards,
Georgi

[NeddySeagoon]

st834,

Code: Select all

Optional: Updating the @world set

Updating the system's @world set is optional and will be unlikely to perform functional changes unless one or more of the following optional steps have been performed:

    A profile target different from the stage file has been selected.
    Additional USE flags have been set for installed packages.

Readers who are performing an 'install Gentoo speed run' may safely skip @world set updates until after their system has rebooted into the new Gentoo environment. 

If you didn't change the configuration, the @world update can be delayed until after your reboot.

[st834]

Hi, Georgi, thank you very much for your answer!

I used a stage3 file (there are not many of them) for systemd, desktop, merged-usr and a new profile (there are many of them) 23.0 for systemd, desktop, plasma, merged-usr. I thought it might be necessary to use the old 17.0 profiles, which I would not like.

Thanks!

Best regards,
Sergei

[st834]

Hi, NeddySeagoon, thanks a lot for your answer!

I read this guide. During the installation of a new system, at this stage, I select a new profile, USE flags and another variables. Here a set of packages is formed, which will be installed and used in the system.

Understand I correctly that in this case, as before, updating the @world (

Code: Select all

emerge --ask --verbose --update --deep --newuse @world

) set is mandatory?

Thanks!

[NeddySeagoon]

st834,

A new install starts with a stage3 tarball. Stage3 tarballs exist for several different profiles. The profile will be set in the stage3
If its a /17.x/ profile, stay on /17.x/ profiles until your install can boot, then follow the news item for upgrading to /23.0/.

The profile upgrade requires rebuilding the entire install, so if your new install is using the old profile, there is little point in choosing a new profile, setting USE flags and updating @world, just to rebuild all of @world again anyway.

If you have a stage3 with a 23.0 profile, you don't need to do the 17.x to 23.0 upgrade as it's free with the stage3.

Changing from a 17.x to a 23.0 profile is a lot more than selecting the new profile

[apiaio]

A new install starts with a stage3 tarball. Stage3 tarballs exist for several different profiles.

I have decided not to upgrade to 23.0 profile but to make the new installation on the same volume on my PC.
How could I find a stage3 tarball with 23.0 profile?

[pietinger]

A new install starts with a stage3 tarball. Stage3 tarballs exist for several different profiles.

I have decided not to upgrade to 23.0 profile but to make the new installation on the same volume on my PC.
How could I find a stage3 tarball with 23.0 profile?

AFAIK it is included in our yesterdays uploads: https://www.gentoo.org/downloads/

[NeddySeagoon]

-releng confirmed that (for amd64) stage3 tarballs with a date stamp 24-March-2024 and later are 23.0 profile based.

[st834]

Thank you very much for the answers. Today I will test the new tarball, and write about the results here.

Best regards

[st834]

Hi, yesterday, as I wrote here, I tested a new profile. The results are not optimistic for me. Installation steps before installation the kernel went as usual and without problems.

When compiling the kernel, I performed the making steps:

Code: Select all

emerge --ask sys-kernel/linux-firmware
emerge --ask sys-kernel/gentoo-sources

emerge --ask sys-kernel/installkernel

make 
make install
make modules_install

All steps were completed successfully except

Code: Select all

make install

Here, instead of three files (kernel-6.6.21-gentoo, System.map, config-6-6-21-gentoo) as before, I receive in /boot only one kernel-6.6.21-gentoo and many information messages on the entire screen. I tried to redirect it to a file to show it here, but these messages are not redirected to the file. In the file I have only

Code: Select all

INSTALL/boot
+depmod 6.6.21-gentoo

Attempts to reinstall sys-kernell/installkernel with the USE systemd-boot flag (as it is written in the news) entails reinstalling systemd with the

Code: Select all

boot

flag, which leads to an error loading the dev-python/puelftools dependency. This is my actual state now. The kernel is compiled but not installed.

At the same time, I take only standard steps at my test installation, shown in the Handbook as mandatory. Previously, they always led to a positive result. I have only two differences this time from previous successful installations:

- profile 23.0
- skipping step

Code: Select all

emerge --ask --verbose --update --deep --newuse @world

after selecting a new profile (this action is now marked as optional in the manual).

I would be grateful to any answers that I do not work correctly, so as to bring the installation with the new profile into working state.

Thanks!

[grknight]

Hi, yesterday, as I wrote here, I tested a new profile. The results are not optimistic for me. Installation steps before installation the kernel went as usual and without problems.

When compiling the kernel, I performed the making steps:

Code: Select all

emerge --ask sys-kernel/linux-firmware
emerge --ask sys-kernel/gentoo-sources

emerge --ask sys-kernel/installkernel

make 
make install
make modules_install

All steps were completed successfully except

Code: Select all

make install

Here, instead of three files (kernel-6.6.21-gentoo, System.map, config-6-6-21-gentoo) as before, I receive in /boot only one kernel-6.6.21-gentoo and many information messages on the entire screen. I tried to redirect it to a file to show it here, but these messages are not redirected to the file. In the file I have only

Code: Select all

INSTALL/boot
+depmod 6.6.21-gentoo

The profile has nothing to do with any of this. If you want the old behavior of installkernel, then disable all of the USE on sys-kernel/installkernel including systemd and systemd-boot.

[st834]

Thank you for your reply.

But it is normal, from the point of view of the new behavior of installkernel, that only the kernel file will go into /boot, System.map is assembled, but remains in the lyrectory of the kernel sources? I could not rebuild the packages (installkernel with systemd-boot, systemd with boot) with the new flags they were disabled initially.

[grknight]

Thank you for your reply.

But it is normal, from the point of view of the new behavior of installkernel, that only the kernel file will go into /boot, System.map is assembled, but remains in the lyrectory of the kernel sources?

With the systemd USE on sys-kernel/installkernel, installkernel bypasses what it has done in the past and just calls systemd's kernel-install. I have no idea what systemd does at that point as I don't use it. I was just noting that it wasn't what your were expecting and what is likely the cause of it.

There is an easy test to see if this is the case. From the kernel directory, run:

Code: Select all

SYSTEMD_KERNEL_INSTALL=0 make install

[st834]

I understand, thanks a lot for the answer. Today I will try to go further and look, how the grub will react to this installation.

[st834]

I tried yesterday, the bootloader was installed, the system started. Thanks so much for your help. Now it remains to read in detail an article about the features of installkernel.

Thanks a lot!