[gentoo-user] Boot has no space left.

[Guillermo García]

Hello,

I was updating my gentoo install, and now it says that /boot has 0
bytes, however i don't know why since i never save anything in that folder.

Any idea on what can i do?

Thanks.

[Dale]

I'd start by doing this:

du -shc /boot/* | sort -h

That will show the size of files in /boot and sort them from smallest to
largest. It could be that you have a large number of kernels and maybe
init thingys in there.  If so, some house cleaning may be required. 

Dale

:-)  :-) 

[Michael]

Also check the output of mount, or findmnt, to determine if /boot is the
mountpoint of a partition and if it is currently mounted or not.

The /boot directory is typically where the boot manager files are installed, as
well as the kernel & initrd images, kernel config and System.map files.

[tastytea]

app-admin/eclean-kernel can help with automating the house cleaning. 😉

[wk...@op.pl]

Dnia 2022-06-27, o godz. 10:35:49
Guillermo García <guillermo18...@gmail.com> napisał(a):

Hello,

Check if there are old kernels (posibly along with initramfs if you use
them). After a while (especially if your /boot partition is not that
big or you use testing kernels) thay can add up to quite some
megabytes.

Hope that helps

--
xWK

[Julien Roy]

On 6/30/22 12:19, Guillermo García wrote:
> Hello guys,
>
> Sorry for not replying the first email i sent, however i got covid and i
> cannot reply you, i forgot the original mail i sent here so i make a new
> one.

If you don't have the replies in your mailbox, you can find them on the
mailing list archive:
https://archives.gentoo.org/gentoo-user/message/ba92573a301b482184cdfe790704110c

Regards,
Julien

[Nikos Chantziaras]

On 30/06/2022 20:11, Guillermo wrote:
> [screenshot]

Doesn't "emerge -a --depclean" remove all these old kernels?

[Guillermo]

Hello,

I still have the same problem, but the command worked fine.

[Michael]

On Thursday, 30 June 2022 19:15:33 BST Guillermo wrote:
> Hello,
>
> I still have the same problem, but the command worked fine.

The command "emerge -a --depclean" will only remove uninstall the kernel
packages, but will not remove files from /usr/src/, or old kernel images and
files from /boot/.

Your /boot partition is full with old kernels you probably no longer use or
need. You have to remove them manually as part of your regular maintenance of
your installation, or you can install and use 'app-admin/eclean-kernel' as
mentioned in the previous thread, to partly automate the cleanup process of
stale kernels. Then update your GRUB to refresh the boot menu.

[Wols Lists]

On 30/06/2022 19:23, Michael wrote:
> On Thursday, 30 June 2022 19:15:33 BST Guillermo wrote:
>> Hello,
>>
>> I still have the same problem, but the command worked fine.
> The command "emerge -a --depclean" will only remove uninstall the kernel

> packages, but will not remove files from/usr/src/, or old kernel images and
> files from/boot/.

As far as I'm aware, depclean only installs files it installed, so it
leaves quite a lot of garbage lying around from kernels, including the
/usr/src/kernel-xx-xx-xx directory and various files involved in making
your kernel, that you've modified.

Cheers,
Wol

[Lee]

The OP should read the section of the Gentoo manual on kernel install to learn what files are installed where. Yea, but just rm the kernels and initramfs's from /boot and you're golden. FWIW, I usually only upgrade my kernel when it's a major revision.

--

Lee 😎 

<ny6...@gmail.com>

[William Kenworthy]

and don't forget to run "uname -a" to get your currently running kernel version and make sure you don't delete that!

"IF" "uname -a" isn't the latest version you have in /boot, some more investigation as to why will be needed.

BillK

[Dale]

William Kenworthy wrote:
>
> and don't forget to run "uname -a" to get your currently running
> kernel version and make sure you don't delete that!
>
> "IF" "uname -a" isn't the latest version you have in /boot, some more
> investigation as to why will be needed.
>
> BillK
>
>

Just to add another method.  I have uprecords installed here.  It lists
the kernels and their uptime.  I keep the last two with reasonably high
uptimes with fairly recent version and the most recent kernel.  I don't
upgrade automatically so I control what and when I update.  Of course, I
also have long uptimes as well.  My thinking on this.  I want kernels
that are known to be stable that I can use as a backup boot option but I
also want newer kernels that have fixes etc in them.  By keeping a
couple with long uptimes, I get stable kernels.  By also picking a
recent kernel version, I get a kernel that I can boot into to see if it
is stable.  Over time, the versions get higher on both parts.  When I do
my checks, I look for kernels with at least 30 days or more of uptime. 
Generally, if a kernel can run that length of time, it is pretty
stable.  That said, I have some with many months of uptime.

When I upgrade to a new kernel, I run for a month or so and then
manually clean out /boot, that would include kernel, init thingy,
System.map and config files. 

Seeing this reminds me it might be a good time to look into updating,
even tho I might not reboot for a while yet. 

Just a thought. 

Dale

:-)  :-) 

[Wols Lists]

On 01/07/2022 00:21, Dale wrote:
> When I upgrade to a new kernel, I run for a month or so and then
> manually clean out /boot, that would include kernel, init thingy,
> System.map and config files.
>
> Seeing this reminds me it might be a good time to look into updating,
> even tho I might not reboot for a while yet.

When I update, I wait until I'm happy the new one seems okay, and then I
just leave the most recent one and the one before.

That said, I need to upgrade, and I need to see if my random hangs are
fixed (there's apparently a bug in the Ryzen 3000, and I'm guessing
that's what I'm hitting).

Cheers,
Wol

[Dale]

After my previous reply, I updated to a newer kernel.  It's in /boot but
it may be months before I reboot.  Anyway, I currently have four kernels
in /boot.  My current running kernel and two backup kernels plus the new
untested one.  Whenever I get around to rebooting and the new kernel
works fine, I'll remove the oldest one including sources etc. 

I try to keep at least two backup kernels.  One reason I do that, the
init thingy.  I admit dracut is working well for me but given the
history I have with those thingys, I want extra protection.  The odds of
three boot options going bad are pretty slim and if it did happen, I
likely have a serious hard drive problem anyway, file system at the very
least.  Either way, I have a lot to worry about. 

Maybe one of the suggestions mentioned here will help the OP.  It seems
he is letting the updater do the install or something and the kernel is
a fast moving target.  One has to have some way, automated or manual, to
clean up the unneeded bits.  I doubt most anyone makes their /boot to
large anyway.  Usually 300 or 400MBs is enough. 

Dale

:-)  :-)