Bug#1034672: new ISO

[Thomas Lange]

Hi,

there are new FAI ISO available using FAI 6.2.2. Does this problem still exist?
-- 
viele Grüße Thomas

Bug#1072538: dracut-core: systemd-cryptsetup missing from generated image

[Thomas Lange]

You are right. It seems that the pacakge does not include the new
dracut module systemd-cryptsetup which handles this.

I will prepare a new version soon.

-- 
regards Thomas

Bug#1067517: solved?

[Thomas Lange]

Hi Florian,

did my explanation helped and solved the problem?
-- 
regards Thomas

Bug#1068250: Switch to 'ng', but calling it 'dracut', don't add 'ng'?

[Thomas Lange]

Yes, I'm thinking about how to switch to dracut-ng.
I've compared the commits of the old and new upstream repo and they
are identical up to the time of the fork.
I hope that I can just replace the upstream repo with the new one, but
I still have to check if that works.

Keeping the old package name dracut is also a good idea.
-- 
regards Thomas

Bug#1071739: marked as done (packages.debian.org: Removal of spam domain from download mirror page)

[Thomas Lange]

> On Fri, 24 May 2024 20:02:35 +0200, Holger Wansing  
> said:

> Hi Thomas,
> you fixed this in master branch.
> Are you sure about this?
> I somehow seem to remember, that debian-master branch is used for 
packages.d.o ...
you are right, debian-master seems to be the correct branch.
I will fix it also in debian-master.

Do you know if the branch master is used for anything?

-- 
 Thomas

Bug#1068250: dracut: Consider switching to the fork dracut-ng

[Thomas Lange]

Yes, I already got this information.

I think I will also prepare dracut-ng to Debian. It then has to go
through the NEW queue. And currently I don't know, when I will start
this.


> On Wed, 22 May 2024 20:54:38 +0200, Evgeni Golov  said:

> FWIW, Fedora switched to this fork starting with Fedora 40 [1].
> [1] https://src.fedoraproject.org/rpms/dracut/

-- 
viele Grüße Thomas

Bug#1071565: autopkgtest: ERROR: cloud-efi.raw is too small: 131 MB. Should be greater 890 MB

[Thomas Lange]

The problem was, that a package could not be downloaded:

833s fai.log:W: Couldn't download package libip4tc2 (ver 1.8.9-2 arch
amd64) at
http://deb.debian.org/debian/pool/main/i/iptables/libip4tc2_1.8.9-2_amd64.deb

Another test run passed without any errors. See 46897030
https://ci.debian.net/packages/f/fai/testing/amd64/

Does this solve your problem? Can we close this bug now?


-- 
 Thomas

Bug#924139: transistion finished

[Thomas Lange]

Hi,

I've remove the old python2 scripts in english/mirror/timestamps/
The only remaining python2 script is now urlcheck.py in the cron
repository, which should not be used any more. Read TODO i nthis
directory.


Therefore I see no blockers for upgrading www-master to bookworm.

Any thoughts?

-- 
regards Thomas

Bug#1071278: systemd 256 breaks dracut

[Thomas Lange]

Hi Luca,

it breaks the current version in unstable and earlier. So please add
Breaks: dracut (<= 060+5-7)

-- 

regards Thomas

Bug#1071182: dracut: requires changes for systemd 256; boot fails otherwise

[Thomas Lange]

The related systemd bug is #1071278
-- 
regards Thomas

Bug#1071182: dracut: requires changes for systemd 256; boot fails otherwise

[Thomas Lange]

I also filed a bug against systemd because this problem can be solved
by both packages and there are plans to replace dracut by dracut-ng.
But that will need more time.


regards Thomas

Bug#1071278: systemd 256 breaks dracut

[Thomas Lange]

Package: systemd
Version: 256~rc2-3
Severity: serious

systemd changed it's behaviour and now makes /usr read-only in the
initrd. This breaks dracut and vice versa.
This bug is releated to #1071182 which says dracut breaks systemd.
Please add a Breaks: dracut(<<..)

Currently I do not know when I will update dracut, because there are
also plans to replace dracut by dracut-ng, which may involve more
time. I not sure in which package I will invest my available time.

In order to not break the systems of our users, IMO the smalles change
would be to add the Breaks: line to systemd.

-- 
 Thomas

Bug#1069967: fai-client: install_packages E: Unable to locate package ...

[Thomas Lange]

Hi Markus,

we had function this in the past. The subroutine mkpackagelist() in
install_packages did this. It was removed, because it could not deal
with variables in package names. So I need to reimplement it.

-- 
viele Grüße Thomas

Bug#872944: #872944 www.debian.org: Remove JavaScript from Policy Manual published on web mirrors

[Thomas Lange]

Currently we have a working solution using js and providing multi page
html. That's a good solution which is already available.


> I did not go deeper into this scenario, I just found 
> https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=877337
> which includes a forward-backword-forward dance switching multiple
> times between multi-page and single-page html variant requests.
A single page html may be an additional option but there's already the
single page txt version and the PDF. That's sufficient and I see no
need in providing more formats of this manual.

Therefore we can close this and I will close 877337.

-- 
regards Thomas

Bug#872944: #872944 www.debian.org: Remove JavaScript from Policy Manual published on web mirrors

[Thomas Lange]

> On Wed, 10 Apr 2024 21:33:50 +0200, Holger Wansing  
> said:


> The second javascript functionality is the full-text search.

> Please note, that I made use of javascript by intend, despite of this bug 
> requesting to remove all js functionality.
Hi holger,

in the past we tried to avoid javascript, but that's long time ago
(like 7 years) and nowadays I see no reason to do web pages without it
if we loose functionality.

So please go ahead and use js. I think the search function is very important.

I think it's important not to load js code on our pages from an
external URL, but to provide it from our web servers (selfhosted).

> Please note, that this decision is not only for debian-policy, but for
> all sphinx-based manuals on Debian website.
> (I hope we don't make different decisions on this question for the 
> various manuals we have. That would make the implentation once again
> more difficult.)
All sphinx-based manuals can use js from my point of view. There's no
reason to handle some manuals differently.


> What should be done now?
Close the bug that request to remove all js functionality (#876241).

-- 
regards Thomas

Bug#987943: www.debian.org: Developers Reference: Sphinx search non-functional: searchindex.js missing

[Thomas Lange]

> On Mon, 8 Apr 2024 00:35:46 +0200, Holger Wansing  
> said:

>> Can be viewed at 
>> 

>> (also with a different html theme, BTW)
>> 
The search works for me. Thanks a lot.

-- 
regards Thomas

Bug#1064593: issue with Debian-style html theme for sphinx-based documents

[Thomas Lange]

Hi Holger,

even if things get more complex, this is a working solution. I'm very
happy for that and there's no need for spending more time into looking
for a perfect solution.
>From my side you get a thank you very much and a GO for applying this patch.

> On Tue, 2 Apr 2024 14:47:12 +0200, Holger Wansing  
> said:

> The 1ftpfiles and 7doc scripts, which need to be adapted for that, and
> also the situation on the www mirrors is getting more complex, so I'm 
unsure 
> if we want this.
> See my patch.

> On the other side, I don't see any other solution apart from developing
> a new theme.

-- 
regards Thomas

Bug#1068250: dracut: Consider switching to the fork dracut-ng

[Thomas Lange]

> On Tue, 02 Apr 2024 19:59:57 +0200, Jörg Behrmann 
>  said:


> Activity in dracut upstream has died down recently and though there is a 
version
> 60 of dracut in sid, upstream has not tagged such a release.
The upstream release process changed, so they do not tag the releases
any more.


> A fork of dracut has been started by some dracut developers at
> https://github.com/dracut-ng/dracut-ng
> It should be evaluated to switch to this fork.
Thanks for the info. I will carefully watch the work on this fork.

-- 
regards Thomas

Bug#1067649: Verification page is not accessible from the homepage

[Thomas Lange]

> So here I adapt my original request: please add the link to /verify at
> the /distrib page; and if possible, also consider renaming the link of
I've added the link now to the /distrib web page.

-- 
regards Thomas

Bug#1067517: AW: Re: Bug#1067517: found bug in checkdisk

[Thomas Lange]

> On Tue, 26 Mar 2024 13:36:44 +, Florian Goth 
>  said:

> OK, on my system, in a fai-nfsroot for bookworm
> the group of /dev/sda was root, and this was also returned by this stat 
command.
This may happen if you do not use systemd inside the nfsroot for
bookworm with FAI 6.2 and newer. You may need to adjust the
/etc/fai/NFSROOT config if you are still using an old version.
In the past this config disabled systemd inside the nfsroot.

Bug#1067517: found bug in checkdisk

[Thomas Lange]

> On Mon, 25 Mar 2024 11:57:10 +, Florian Goth 
>  said:

> stat -c %G
> returns the group of the owner, but nothing related to a device type.
And this group is "disk" in case it's a disk.

-- 
viele Grüße Thomas

Bug#1067649: Verification page is not accessible from the homepage

[Thomas Lange]

I don't think we need the link on the startpage, but maybe on the
/distrib page where we provide other ISO downloads.

But in the end I try to remember whenever I have verified an ISO by
myself.

I can't remember and I guess I never did it. Maybe I
check that I'm using https://www.debian.org. That's all I need to
trust.

-- 
regards Thomas

Bug#1042411: closed by Thomas Lange (found problem)

[Thomas Lange]

Hi Marc,

you are right. I've reopened this bug and will try to fix it in the
next version.

-- 
viele Grüße Thomas

Bug#1064441: does not copy code

[Thomas Lange]

Package: fai-client
Severity: normal

Since qemu-user-static without the F flag is not supported on Debian
anymore the binaries are never copied. Therefore lowering the severity.
This may only have happended until Debian buster.

-- 
viele Grüße Thomas

Bug#1064442: fai-client - Unable to disable mkramdisk

[Thomas Lange]

This is the very easy fix:

 # set default dirs if FAI_RAMDISKS not already set
-: ${FAI_RAMDISKS:="$target/var/lib/dpkg"}
+: ${FAI_RAMDISKS="$target/var/lib/dpkg"}


@wwaldi: Do you think it's still wrong?

-- 
regards Thomas

Bug#949610: will it ever work?

[Thomas Lange]

Hi,

currently I think it's not possible to have raid+luks+lvm prepared
with setup-storage.

I didn't manage to prepare a working example.
-- 
regards Thomas

Bug#1050383: reverting the fix

[Thomas Lange]

Hi,

I will revert this fix in FAI 6.2.1, because it broke other
things. Our LVM on soft RAID example does not work any more.

You application of using a hook for creating a md device and then set
disklist=/dev/md is not what setup-storage supports.
Why not use your heuristic and create then the disklist (of plain
disks) and let setup-storage create the md devices?

I'm not sure if we can fix your problem without breaking the normal
configs.

-- 
regards Thomas

Bug#968122: I confirm this bug

[Thomas Lange]

I can confirm this bug.

Using qemu/kvm with autopkgtests I have
keyboard-configuration 1.221 inside the testbed and the postinst
script of keyboard-configuration fails with

cat: '/sys/bus/usb/devices/*:*/bInterfaceClass': No such file or directory
cat: '/sys/bus/usb/devices/*:*/bInterfaceSubClass': No such file or directory
cat: '/sys/bus/usb/devices/*:*/bInterfaceProtocol': No such file or directory

-- 
regards Thomas

Bug#1063146: Problem with setup-storage

[Thomas Lange]

Here's the patch

https://github.com/faiproject/fai/commit/cd33235318279a96267d3377661dbeba93eacac3

-- 
viele Grüße Thomas

Bug#1063146: fai-setup-storage: New dependency on /usr/lib/fai/subroutines

[Thomas Lange]

I can add a Depends or Recommends on fai-client for the package
fai-setup-storage. What do you think is better?

-- 
regards Thomas

Bug#1063146: fai-setup-storage: New dependency on /usr/lib/fai/subroutines

[Thomas Lange]

Hi Stephen,

thanks for the report. setup-storage does not need fai-client if you
use the option -D and specify the list of disks.

So it's not always needed. But I will prepare a patch for it.

-- 
regards Thomas

Bug#1017039: (no subject)

[Thomas Lange]

Sorry, but maybe I missread your email.

If this change is applied upstream, then Debian will include the change
with the next Debian release.
It's a chang to 90overlayfs and not 90overlay-root which is Debian
specific and which is needed for my FAI software. I just want to make
sure that 90overlay-root will not be broken by some patches.

If 90overlayfs works with nfs, I may drop the 90overlay-root modules
in the future after I made some changes in the FAI (Fully Automatic
Installation) software.

-- 
regards Thomas

Bug#1017039: (no subject)

[Thomas Lange]

We have a working 90overlay-root module in Debian which can mount a
NFS directory. Currently I see no need to replace this with a new
implementation, that was not working when I tested it last time. I
also wonder why upstream does not include it? Are there quality issues?

Currently I have no time to spend more time into it, because I have to
focus on my other software projects.
It would be nice if you could test and report if it works for you.
-- 
regards Thomas

Bug#1057807: dracut-network: broken on sid ("Module 'ifcfg' cannot be found.", and /etc/dracut.conf.d/11-ifcfg.conf add_dracutmodules+=" ifcfg " but modules.d/45ifcfg gone vs bookworm)

[Thomas Lange]

A new dracut version 060+5-1 will be available soon.
060+5 is the same as 059+212 because upstream changed its release process.

-- 
reagrds Thomas

Bug#1057807: dracut-network: broken on sid ("Module 'ifcfg' cannot be found.", and /etc/dracut.conf.d/11-ifcfg.conf add_dracutmodules+=" ifcfg " but modules.d/45ifcfg gone vs bookworm)

[Thomas Lange]

I guess I can fix this by adding

remove-on-upgrade /etc/dracut.conf.d/11-ifcfg.conf
to debian/dracut-network.conffiles

--
Thomas

Bug#1057284: migrate get-www-stats to python3

[Thomas Lange]

Package: www.debian.org

The script get-www-stats is using python2. It does not work with
python3 because it uses counts.iteritems which must be migrated.

The script should also ignore log entries with produces 403 on the web
server. Currently this is half of them because of a lot of access to
the uncompressed oval...xml files.

We only have this python2 script in the webwml repo. Python scripts in
security/oval are already python3.

regards Thomas

Bug#1056151: reason

[Thomas Lange]

The comment in lib/updatebase explains why the diversion for
/sbin/init was added:

# if libc is upgraded init u is called in chroot environment and
# then init will eat up much cpu time

This was a rare case in the past and maybe we can remove this
diversion? What do you think?

-- 
 Thomas

Bug#1029185: moreinfo

[Thomas Lange]

Hi Reiner,

did you tried to convert the .deb to .rpm? Does it work for you?
-- 
 Thomas

Bug#1033040: very simple to create a different format

[Thomas Lange]

Hi,

if you want to create a .vdi image just call

# fai-diskimage . image.vdi

What else is needed as an example?

-- 
regards Thomas

Bug#1039194: not needed

[Thomas Lange]

FYI: Since the next FAI version will move to systemd I will not add an
override.

-- 
 Thomas

Bug#1050383: fix applied

[Thomas Lange]

I've check with the upcoming FAI version, and there the md device
names do start with md0.

I've already added your fix to the git repo.
-- 
regards Thomas

Bug#1056151: systemd will be used in the future

[Thomas Lange]

Hi Mika,

the additional test before calling fai-divert was added because in
the next FAI release, we will move from sysv to systemd by default.

I think I will not drop the code for sysv support immediately.

-- 
viele Grüße Thomas

Bug#1056380: Acknowledgement (reprepro hangs on Ubuntu package, unzstd process waiting)

[Thomas Lange]

The same bug can be found in 5.4.2-1 on a bookworm system.

-- 
 Thomas

Bug#1056380: Acknowledgement (reprepro hangs on Ubuntu package, unzstd process waiting)

[Thomas Lange]

This bug may be related to #1050321

-- 
 Thomas

Bug#1056380: reprepro hangs on Ubuntu package, unzstd process waiting

[Thomas Lange]

Package: reprepro
Version: 5.3.1-1+deb12u1

reprepro hangs when adding one package, but it does not hang when
reprepro runs under strace. Also it does not hang with every package.
When it hangs, I see unzstd process not finishing.

I'm using 5.3.1-1+deb12u1 on a bookworm system. I have a small
directory containing:

./grub-pc_2.06-2ubuntu7.2_amd64.deb
./M
./M/conf
./M/conf/distributions

$ file grub-pc_2.06-2ubuntu7.2_amd64.deb
grub-pc_2.06-2ubuntu7.2_amd64.deb: Debian binary package (format 2.0), with 
control.tar.zs, data compression zst

$ cat M/conf/distributions
Codename: cskoeln
Architectures: amd64
Components: main
DebIndices: Packages Release .
Description: Package repository created by fai-mirror
Label: fai-project.org
Origin: fai-mirror


[~]$ reprepro -V --ignore=extension -b M includedeb cskoeln 
grub-pc_2.06-2ubuntu7.2_amd64.deb
Created directory "M/db"



In another teminal I can see
$ pgrep -a unzst
47472 /usr/bin/unzstd

$ pstree -p
───bash(42341)───reprepro(47461)───unzstd(47472)─┬─{unzstd}(47473)
 └─{unzstd}(47474)


$ find -ls
 36880744  4 drwxr-xr-x   3 langeuser  4096 Nov 22 02:57 .
 36880795132 -rw-r--r--   1 langeuser131820 Nov 22 02:57 
./grub-pc_2.06-2ubuntu7.2_amd64.deb
 36880796  4 drwxr-xr-x   4 langeuser  4096 Nov 22 03:02 ./M
 36880763  4 drwxr-xr-x   2 langeuser  4096 Nov 22 03:02 ./M/db
 36880808 64 -rw-r--r--   1 langeuser 65536 Nov 22 03:02 
./M/db/contents.cache.db
 36880803 64 -rw-r--r--   1 langeuser 65536 Nov 22 03:02 
./M/db/packages.db
 36880805 64 -rw-r--r--   1 langeuser 65536 Nov 22 03:02 
./M/db/references.db
 36880801  0 -rw---   1 langeuser 0 Nov 22 03:02 
./M/db/lockfile
 36880806 64 -rw-r--r--   1 langeuser 65536 Nov 22 03:02 
./M/db/checksums.db
 36880804  4 -rw-r--r--   1 langeuser31 Nov 22 03:02 
./M/db/version
 36880797  4 drwxr-xr-x   2 langeuser  4096 Nov 22 02:56 ./M/conf
 36880798  4 -rw-r--r--   1 langeuser   184 Nov 22 02:56 
./M/conf/distributions


Now I interrupt the reprepo call. After typing ^C I see

Interruption in progress, interrupt again to force-stop it (and risking 
database corruption!)

grub-pc_2.06-2ubuntu7.2_amd64.deb: component guessed as 'main'
Created directory "M/pool"
Created directory "M/pool/main"
Created directory "M/pool/main/g"
Created directory "M/pool/main/g/grub2"
There have been errors!

faime[~]$ find
.
./grub-pc_2.06-2ubuntu7.2_amd64.deb
./M
./M/db
./M/db/contents.cache.db
./M/db/packages.db
./M/db/references.db
./M/db/checksums.db
./M/db/version
./M/pool
./M/pool/main
./M/pool/main/g
./M/pool/main/g/grub2
./M/pool/main/g/grub2/grub-pc_2.06-2ubuntu7.2_amd64.deb
./M/conf
./M/conf/distributions

The indices are not created.

When I run the same command under strace it does not hang but finished
without errors. Very strange!


$ strace -o /tmp/lll reprepro --ignore=extension -b M includedeb cskoeln 
grub-pc_2.06-2ubuntu7.2_amd64.deb



Now I install an older version of reprepo from bullseye on my bookworm machine:
ii  reprepro   5.3.1-1~bpo11+1 amd64Debian package repository 
producer


faime[~]$ reprepro -V --ignore=extension -b M includedeb cskoeln 
grub-pc_2.06-2ubuntu7.2_amd64.deb
Created directory "M/db"
zstd: error 70 : Write error : cannot write block : Broken pipe
grub-pc_2.06-2ubuntu7.2_amd64.deb: component guessed as 'main'
Created directory "M/pool"
Created directory "M/pool/main"
Created directory "M/pool/main/g"
Created directory "M/pool/main/g/grub2"
Exporting indices...
Created directory "M/dists"
Created directory "M/dists/cskoeln"
Created directory "M/dists/cskoeln/main"
Created directory "M/dists/cskoeln/main/binary-amd64"


$ find
.
./grub-pc_2.06-2ubuntu7.2_amd64.deb
./M
./M/db
./M/db/contents.cache.db
./M/db/packages.db
./M/db/release.caches.db
./M/db/references.db
./M/db/checksums.db
./M/db/version
./M/dists
./M/dists/cskoeln
./M/dists/cskoeln/Release
./M/dists/cskoeln/main
./M/dists/cskoeln/main/binary-amd64
./M/dists/cskoeln/main/binary-amd64/Release
./M/dists/cskoeln/main/binary-amd64/Packages
./M/pool
./M/pool/main
./M/pool/main/g
./M/pool/main/g/grub2
./M/pool/main/g/grub2/grub-pc_2.06-2ubuntu7.2_amd64.deb
./M/conf
./M/conf/distributions


Although I got an error messages from zstd, it does not hang and
everything looks fine.


-- 
regards Thomas

Bug#1056151: fai-client: Diversion makes /usr/sbin/init vanish in /usr-move conditions

[Thomas Lange]

Hi Chris,

> On Fri, 17 Nov 2023 20:54:29 +0100, Chris Hofstaedtler  
> said:

> Now a problem arises, when:
> - I use a basefile tar.gz, made with an old systemd (say, it uses
>   testing as of today)
> - During baseupdate, systemd gets updated and moves its file (say, I'm
>   actually installing unstable)
Mmm, I do not know if it's possible to catch this.
Using an old system which is then updated may always cause strange
problems.

Does this problem also arises when you use a basefile from unstable?
This is the main question I have.

fai-client 5.11? I do not know this version. I guess it's some
other version.

-- 
viele Grüße Thomas

Bug#1056059: dracut: systemd 255: dracut fails to boot due to lack of systemd-executor

[Thomas Lange]

I've already included the upstream patch to the git master branch of
dracut on salsa. The next dracut release in Debian will include the
fix.

-- 
regards Thomas

Bug#1050383: setup-storage confused by valid md device names

[Thomas Lange]

Hi,

I guess the \d{3,} regex comes from the fact, that during an initial
FAI installation (booting the FAI environment) the md devices are
always higher that 127 IIRC. If you use setup-storage on a normal
Debian system, this is not true any more.

I will think about changing that.
-- 
regards Thomas

Bug#470185: can we close this bug?

[Thomas Lange]

Hi Cord,

you have requested for help in 2008. But it seems that you actively
maintain the package at least in the last 3 years. Please think about
if this bug can be closed.

-- 
viele Grüße Thomas

Bug#1055042: FAI client does not boot to Hard Disk after installation

[Thomas Lange]

Hi,

you have a major network problem.

"Temorary failure resolving 'deb.debian.org'"

That's why a lot of packages (maybe all) are not installed.
Please fix problem, otherwise you will never have a bootable
installation.
For e.g. grub-efi is not installed, and without grub your computer
cannot boot.

-- 
regards Thomas

Bug#1055026: do not create web statistic for all language combinations

[Thomas Lange]

There's also the problem, that a lot of files are outdated.

After I removed some files for german manually I now see this:

https://www.debian.org/devel/website/stats/de.en.html
Last Modified: Sun, Oct 29 15:26:29 UTC 2023   Last Built: Sun, Oct 29 15:29:11 
UTC 2023

This was recreated today, but

https://www.debian.org/devel/website/stats/de.de.html
Zuletzt geändert: Dienstag den 13. Jun 2023 um 19:26:34 Uhr UTC   Zuletzt 
gebaut: Dienstag den 13. Jun 2023 um 22:59:06 Uhr UTC 

Similar for nl.nl.html and nl.de.html.


-- 
 Thomas

Bug#1055028: do not create po statistic for all language combinations

[Thomas Lange]

Package: www.debian.org

Currently we generate several hundereds of files in international/l10n/po/
A lot of them are around 2.2M big.

We have an overall of 8700 files in this directory.
Do we really need statistic like these? For e.g.
/international/l10n/po/de@hebrew
/international/l10n/po/sa.de.html

We should review if we really need all these files in there.

The directory currently needs 2.2 GB disk space.
-- 
regards Thomas

Bug#1055026: do not create web statistic for all language combinations

[Thomas Lange]

Package: www.debian.org


Currently we generate nearly 700 html files in /devel/website/stats
most of them not needed and each around 2.5M big so we have 1.6GB data
there.

The web logs show that only the file in the native language and the
engish version is accessed. There are also no link to the other translations.

For example combinations like de.fr.html, fr.de.html are not needed.
I wonder why these are generated and I guess we could save a lot of
building time in 4webwml.

-- 
regards Thomas

Bug#886792: wrong results

[Thomas Lange]

When search for

You have searched for packages that names contain linux-image in suite(s) 
stable, section(s) main, and all architectures.

https://packages.debian.org/search?keywords=linux-image=names=1=stable=main

I get
Your keyword was too generic.
Please consider using a longer keyword or more keywords.

When I click on " results have not been displayed due to the search..."

I also see packages from oldoldstable.


So it's not possible to just get the list of package names from stable

-- 
regards Thomas

Bug#442083: still valid?

[Thomas Lange]

I wonder if this bug is still valid?

-- 
 Thomas

Bug#977160: new info?

[Thomas Lange]

> I'll think a bit more about what's the right thing to do now...
Hi Joost,

is there an update on that topic?
-- 
regards Thomas

Bug#354432: what should be done now

[Thomas Lange]

Can anyone decide what to do with this bug?
Wontfix, closing or is the patch working?

-- 
regards Thomas

Bug#754601: still a problem?

[Thomas Lange]

Hi victory,

is this still a problem?

-- 
regards Thomas

Bug#772822: link to mail possible

[Thomas Lange]

There's now a new list which includes a direct link to the DSA
announcement mail and the security tracker.
This can also be done for DLAs.

https://www.debian.org/security/dsa.html#DSAS

-- 
regards Thomas

Bug#1054263: faicd64-large_6.0.3.iso "Cannot find grub.cfg"

[Thomas Lange]

Thanks for the screenshots.
I found the problem. It only occurs in an UEFI environment.

If you boot in an EFI environment and your hostname is faiserver, then
the FAI class FAISERVER is defined. In package_config/FAISERVER the
package grub-pc is listed, which conflicts with grub-efi. That's why
no grub is installed and your host cannot boot.

Since you cannot edit the files on the ISO, you can only enter the
grub menu (pressing e) and remove the part
hostname=faiserver

But then some other things may be missing.

-- 
regards Thomas

Bug#1054263: faicd64-large_6.0.3.iso "Cannot find grub.cfg"

[Thomas Lange]

Please provide the log files fai.log and scripts.log of this
installation for better debugging.

-- 
regards Thomas

Bug#861781: update

[Thomas Lange]

I've removed the people not involed any more and the OVAL
organisation.

We may remove all organisations, because I'm not sure if
Debian is involved in any organisations any more.

-- 
 Thomas

Bug#1053896: Acknowledgement (content very outdated)

[Thomas Lange]

It seems that on https://www.debian.org/doc/user-manuals#java-faq we
publish the a version from 2014 from
https://salsa.debian.org/ddp-team/java-faq
which is much more outdated that the version in the package
java-policy.

-- 
regards Thomas

Bug#1053896: content very outdated

[Thomas Lange]

Package: java-policy
Version: 0.57


Hi,

the Java FAQ is very outdated. It only covers i386 and java 6/7 up to
Debian wheezy. There are links to alioth and it talks about
icedtea-7-plugin.

The salsa repo shows that it was not updated since 9 years.

Maybe we should remove the FAQ.
--
regards Thomas

Bug#1053508: sitemap.wml also lists external web pages, review all content

[Thomas Lange]

Package: www.debian.org

The sitemap also includes external links like qa.debian.org,
lists.d.o, even wiki.d.o But a sitemap should only list internal pages.

Also the selection seems to be very random. Why should archives of
mailing lists listed here? Is devel/developers.loc that important to
be listed on a sitemap?

Why are these listed?
- Procedures for submitting a General Resolution proposal or amendment
- Standard Resolution Procedure
It's sufficient to list a link to our vote page


Please review all content. Less is more!
-- 
regards Thomas

Bug#548024: more info, still an issue?

[Thomas Lange]

I want to know if this still is an issue (bug was opened 10+ years
ago), or did we missed to close the bug. I want to avoid that we still
need to ignore it for all the next releases.

-- 
regards Thomas

Bug#983079: blends bug

[Thomas Lange]

Hi Andreas,

can you take care about this bug?
-- 
viele Grüße Thomas

Bug#1053445: Merge request regarding 'Please migrate Release Notes to reStructuredText'

[Thomas Lange]

Hi Holger,

I really like the idea no to produce release notes for each
architecture but only one. Moving to sphinx is also nice.

Sorry, if I broke your MR, by adding code that checks if something
changed in the git repo. I think I can easily add this to your code
later. So maybe we copy your version of 7release-notes and after that
I add my code.

Do you know how long the build process takes using sphinx? I've added
the code, because the build took around 90 minutes using docbook.

Any other things I should keep an eye on?
-- 
regards Thomas

Bug#905440: add info

[Thomas Lange]

The tool is already installed in www-master.
We can use

$ git restore-mtime

A test of a git restore-mtime run on www-master took one minute.
We still have to add this to the cron jobs.

--
regards Thomas

Bug#720989: closing

[Thomas Lange]

Hi,

I think in the meantime we improved the situation a lot.
On our startpage we now have a direct link to an ISO (we do no call
it CD image any more) and we have a link to other images
(this is https://www.debian.org/distrib/) which is much cleaner
than before. It would be nice to update the german translations
https://www.debian.org/distrib/index.de.html
but then this bug should be closed.

regards Thomas

Bug#680109: did we solve a problem?

[Thomas Lange]

Hi,

I wonder if there's really a big problem we try to solve here?
Do we have more complaints since 2012?
We had a longer discussion here, but still no solution yet?

If there are not that many complaints, we should just close this bug.

regards Thomas

Bug#863579: how to proceed

[Thomas Lange]

What can we do now to solve this bug?

Is there an active wiki team?
If someone needs the contact infos, using a search engine with "debian
wiki contact" will directly find the correct page. So IMO no need to
add them to the into/organisation page.

Any comments?

Bug#859123: new DAS list

[Thomas Lange]

We now have an automatically generated list of DSA, which does not need
any manual adding the infos. See
https://www.debian.org/security/dsa.html#DSAS

This list also links to the security tracker and the security
announcement mail.

My plan is to remove the need of manual editing wml and data files for
every DSA/DLA.

regards Thomas

Bug#1034672: Please check with new ISO

[Thomas Lange]

Hi,

please check if ventoy has this problem exists with a current ISO
for e.g. https://fai-project.org/fai-cd/faicd64-large_6.0.3.iso

-- 
regards Thomas

Bug#1039105: I can confirm this bug, fix works for me

[Thomas Lange]

Hi Dirk,

I'm a vm user for reading my mail.
Thanks a lot for this bug report and fix. It also works for me.
Patching /etc/emacs/site-start.d/50vm-init.el still compiles vm-vars
and vm-version, but I can start the VM mode and read my mails with
emacs 28 using bookworm.

After this first test, I can upgrade my computer to bookworm. Yeah!

I vote for adding this fix into a new vm version for bookworm proposed
updates.

-- 
reagrds Thomas

Bug#1039019: bookworm-pu: package fai/6.0.3+deb12u1

[Thomas Lange]

Package: release.debian.org
Severity: normal
Tags: bookworm
User: release.debian@packages.debian.org
Usertags: pu
X-Debbugs-Cc: f...@packages.debian.org
Control: affects -1 + src:fai


This upload fixes #1037329


[ Reason ]
A change from FAI 5.10.3 (in bullseye) to 6.0 (currently 6.0.3 in bookworm)
removed some code which sets the lifetime for IP addresses to forever. This 
change must be reversed.

[ Impact ]
A network installation hangs completly after the lease time of the IP address 
expires.
FAI does not run a dhclient, because it uses a nfsroot.

[ Tests ]
Since we used this code years before, and I only reverted the commit which 
removes the code, no tests are needed.

[ Risks ]
No risks,.

[ Checklist ]
  [X ] *all* changes are documented in the d/changelog
  [X ] I reviewed all changes and I approve them
  [X ] attach debdiff against the package in (old)stable
  [X] the issue is verified as fixed in unstable

[ Changes ]
For each network interface, set IP address filetime to forever.


 bin/fai  |6 ++
 debian/changelog |6 ++
 2 files changed, 12 insertions(+)

diff -Nru fai-6.0.3/bin/fai fai-6.0.3+deb12u1/bin/fai
--- fai-6.0.3/bin/fai   2023-01-12 10:22:03.0 +0100
+++ fai-6.0.3+deb12u1/bin/fai   2023-06-24 12:57:09.0 +0200
@@ -126,6 +126,12 @@
 
 cat /proc/kmsg >/dev/tty4 &
 
+# fix IP address lifetime
+   ip -4 -br a | awk '/UP / {if ($3) print $3 " " $1}' | \
+while read addr iface; do
+ip -4 addr change "$addr" dev "$iface" valid_lft forever 
preferred_lft forever
+done
+
 # enable EFI variables
 if [ -d /sys/firmware/efi ]; then
mount -t efivarfs none /sys/firmware/efi/efivars
diff -Nru fai-6.0.3/debian/changelog fai-6.0.3+deb12u1/debian/changelog
--- fai-6.0.3/debian/changelog  2023-05-24 11:57:11.0 +0200
+++ fai-6.0.3+deb12u1/debian/changelog  2023-06-24 13:02:26.0 +0200
@@ -1,3 +1,9 @@
+fai (6.0.3+deb12u1) bookworm; urgency=low
+
+  * fai: set IP address lifetime to forever, Closes: #1037329
+
+ -- Thomas Lange   Sat, 24 Jun 2023 13:02:26 +0200
+
 fai (6.0.3) unstable; urgency=high
 
   *  get-boot-info: write $SERVER only if string is non-epmty

Bug#1037488: www.debian.org: drop static/packages-site.css entirely?

[Thomas Lange]

> On Tue, 13 Jun 2023 14:27:34 +0200, Cyril Brulebois  
> said:

> Probably best to finally drop it, to avoid wasting more brain time on
> suite updates (for that particular file)?
Yes, please drop it.

-- 
viele Grüße Thomas

Bug#1037329: IP address gets lost

[Thomas Lange]

Package: fai-client
Version: 6.0
Severity: important

A FAI client booting via network get it's IP address from a DHCP
server and the kernel holds this IP only for the lease time. Since we
do not run dhclient on the client, it looses its IP address after some
time and the network installation hangs. You cannot log into the
machine any more.
Error message on the console:

nfs: server 192.168.33.250 not responding, timed out



The code that sets the lifetime for IP addresses to forever was
removed in aa429361806091b1f2abed8b31f15f6a30eec6e2, because we wanted
to solve a problem by starting dhclient. But that code was removed in
55df9af456e0fc3048d3fefb490fe4c47fe271ae, so we must use the old code.


-- 
regaards Thomas

Bug#1032453: non-free-firmware changes and documentation needs to happen after SC change

[Thomas Lange]

I totally agree that there's no need to modify the Why Debian page.

-- 
regards Thomas

Bug#1030312: avoid aptitude

[Thomas Lange]

Yes, this is a bug.

I suggest to use apt-get instead. Therefore use PACKAGES install or
PACKAGES install-norec.

aptitude is not inside the nfsroot by default. That's why this will be
a low priority for me to fix.

Bug#1036728: unblock: fai/6.0.3

[Thomas Lange]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package fai


I've added a small patch, which prevents overwriting a variable
with an empty value. Now we only write if the variable is not empty.

The PR for this says: do not write an (always empty) SERVER= line to bootlog


 debian/changelog  |7 +++
 lib/get-boot-info |4 +++-
 2 files changed, 10 insertions(+), 1 deletion(-)


[ Reason ]
>From the PR:

when i dug around a little it appears that the problem is in the
setnet function of /usr/lib/fai/get-boot-info. it looks like commit
6e1cbda removed the line in setnet where $SERVER was previously
defined but left in the bit where the (now empty) variable is written
to boot.log (which is consumed later). if i remove that leftover line
fai reports the expected/desired:


[ Impact ]

[ Tests ]
I've reviewed the code change.

[ Risks ]
No risks expected.

[ Checklist ]
  [X] all changes are documented in the d/changelog
  [X] I reviewed all changes and I approve them
  [X] attach debdiff against the package in testing

unblock fai/6.0.3



diff -Nru fai-6.0.2/debian/changelog fai-6.0.3/debian/changelog
--- fai-6.0.2/debian/changelog  2023-05-07 17:25:35.0 +0200
+++ fai-6.0.3/debian/changelog  2023-05-24 11:57:11.0 +0200
@@ -1,3 +1,10 @@
+fai (6.0.3) unstable; urgency=high
+
+  *  get-boot-info: write $SERVER only if string is non-epmty
+ fixes https://github.com/faiproject/fai/pull/118
+
+ -- Thomas Lange   Wed, 24 May 2023 11:57:11 +0200
+
 fai (6.0.2) unstable; urgency=high
 
   [ Thomas Lange ]
diff -Nru fai-6.0.2/lib/get-boot-info fai-6.0.3/lib/get-boot-info
--- fai-6.0.2/lib/get-boot-info 2022-11-13 00:32:00.0 +0100
+++ fai-6.0.3/lib/get-boot-info 2023-05-24 11:56:44.0 +0200
@@ -86,9 +86,11 @@
IPADDR=$IPADDR
CIDR=$CIDR
GATEWAYS=$GATEWAYS
-   SERVER=$SERVER
 EOF
 fi
+if [ -n "$SERVER" ]; then
+   echo "SERVER=$SERVER" >> $bootlog
+fi
 
 if [ -n "$DOMAIN" ]; then
 # DOMAIN was specified on the kernel command line

Bug#1036229: unblock: fai/6.0.2

[Thomas Lange]

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: unblock

Please unblock package fai


In a few configs and scripts the non-free-firmware section was added.
In the script mkdebmirror bullseye was replaced with bookworm.
The script 50-misc can now handle the non-free-firmware section, but
also handle older releases that do not have this section.
See changelog below.


[~]$ debdiff fai_6.0.1.dsc fai_6.0.2.dsc|diffstat
 bin/fai-mirror   |4 ++--
 conf/sources.list|6 +++---
 debian/changelog |   10 ++
 debian/fai-doc.doc-base.package  |3 ---
 debian/rules |3 ---
 doc/Makefile |7 +--
 examples/simple/scripts/LAST/50-misc |8 ++--
 utils/mkdebmirror|6 +++---
 8 files changed, 25 insertions(+), 22 deletions(-)



[ Reason ]
adding non-free-firmware is essential for the users, it's also important to 
support
older Debian releases.

[ Impact ]
(What is the impact for the user if the unblock isn't granted?)

[ Tests ]
I've review the code change and done manual tests of the new code and config.

[ Risks ]
The changes are not that big and mostly affect files in the fai-doc
package. Adding the non-free-firmware section is trivial. Also
removing the generation of postscript documentation is trivial.

[ Checklist ]
  [X] all changes are documented in the d/changelog
  [X] I reviewed all changes and I approve them
  [X] attach debdiff against the package in testing

unblock fai/6.0.2



diff -Nru fai-6.0.1/bin/fai-mirror fai-6.0.2/bin/fai-mirror
--- fai-6.0.1/bin/fai-mirror2022-09-25 12:00:17.0 +0200
+++ fai-6.0.2/bin/fai-mirror2023-05-07 16:29:11.0 +0200
@@ -365,7 +365,7 @@
 cat > $mirrordir/conf/distributions <http://deb.debian.org/debian bookworm main contrib non-free
-deb http://deb.debian.org/debian-security bookworm-security main contrib 
non-free
-#deb http://deb.debian.org/debian bookworm-backports main contrib non-free
+deb http://deb.debian.org/debian bookworm main contrib non-free 
non-free-firmware
+deb http://deb.debian.org/debian-security bookworm-security main contrib 
non-free non-free-firmware
+#deb http://deb.debian.org/debian bookworm-backports main contrib non-free 
non-free-firmware
 
 # repository that may contain newer fai packages for bookworm
 deb http://fai-project.org/download bookworm koeln
diff -Nru fai-6.0.1/debian/changelog fai-6.0.2/debian/changelog
--- fai-6.0.1/debian/changelog  2023-03-08 19:38:00.0 +0100
+++ fai-6.0.2/debian/changelog  2023-05-07 17:25:35.0 +0200
@@ -1,3 +1,13 @@
+fai (6.0.2) unstable; urgency=high
+
+  [ Thomas Lange ]
+  * add support for non-free-firmware, Closes: #103
+  * mkdebmirror: bullseye -> bookworm, Closes: #1035608
+  * doc/Makefile, debian/rules,fai-doc.doc-base.package:
+do not create ps files
+
+ -- Thomas Lange   Sun, 07 May 2023 17:25:35 +0200
+
 fai (6.0.1) unstable; urgency=low
 
   [ Joerg Behrmann ]
diff -Nru fai-6.0.1/debian/fai-doc.doc-base.package 
fai-6.0.2/debian/fai-doc.doc-base.package
--- fai-6.0.1/debian/fai-doc.doc-base.package   2012-05-02 23:05:48.0 
+0200
+++ fai-6.0.2/debian/fai-doc.doc-base.package   2023-05-07 16:56:59.0 
+0200
@@ -5,9 +5,6 @@
  Debian GNU/Linux.
 Section: Debian
 
-Format: postscript
-Files: /usr/share/doc/fai-doc/fai-guide.ps.gz
-
 Format: text
 Files: /usr/share/doc/fai-doc/fai-guide.text.gz
 
diff -Nru fai-6.0.1/debian/rules fai-6.0.2/debian/rules
--- fai-6.0.1/debian/rules  2021-05-01 22:23:53.0 +0200
+++ fai-6.0.2/debian/rules  2023-05-07 16:56:59.0 +0200
@@ -15,9 +15,6 @@
 override_dh_installdocs:
dh_installdocs -Nfai-server -Nfai-quickstart
sed -i 's/FAIVERSIONSTRING/$(VERSIONSTRING)/' 
debian/fai-client/usr/share/doc/fai-client/README
-   # Remove embedded temporary directory for reproducible builds
-   sed -i -e 's,/tmp/tmp.*/fai-guide.ps,fai-guide.ps,g' 
debian/fai-doc/usr/share/doc/fai-doc/fai-guide.ps
-   sed -i -e 's,/tmp/tmp.*/fai-guide.dvi,fai-guide.dvi,g' 
debian/fai-doc/usr/share/doc/fai-doc/fai-guide.ps
 
 override_dh_installchangelogs:
dh_installchangelogs -Nfai-server -Nfai-quickstart
diff -Nru fai-6.0.1/doc/Makefile fai-6.0.2/doc/Makefile
--- fai-6.0.1/doc/Makefile  2022-10-06 19:06:21.0 +0200
+++ fai-6.0.2/doc/Makefile  2023-05-07 16:29:11.0 +0200
@@ -6,7 +6,7 @@
 OPT = --dblatex-opts "-P latex.output.revhistory=0"
 
 
-free:  text html ps pdf
+free:  text html pdf
 #  echo "`grep -c FIXME $(DOC).txt` FIXMEs left to fix:"
 #  grep FIXME $(DOC).txt
 
@@ -21,11 +21,6 @@
a2x $(OPT) -L --icons -a toc -a toclevels=3 -f pdf $(DOC).txt
rm -f $(DOC).xml $(DOC).fo
 
-.NOTPARALLEL:  ps
-ps: $(DOC).txt images
-   a2x $(OPT) -L --icons -a toc -a toclevels=3 -f ps $(DOC).txt
-

Bug#1035608: mkdebmirror still uses bullseye

[Thomas Lange]

Package: fai-doc
Severity: normal

The script mkdebmirror still uses the bullseye release. Please upgrade
to bookworm.

-- 
regards Thomas

Bug#1034672: check your grub.cfg

[Thomas Lange]

Hi,

you did not say which FAI version you are using and how you did create
the bookworm image. I wonder why you still have
  gfxpayload=$resolution
in the grub.cfg? This was removed in FAI 5.9.1.
Also insmod all_video is used since a long time.


Please also check if this ISO does not work with Ventoy
https://fai-project.org/fai-cd/faicd64-large_6.0.iso

-- 
reagrds Thomas

Bug#1035555: FAI examples do not manage non-free-firmware section

[Thomas Lange]

Package: fai-doc
Severity: important


The examples of the configuration scripts do not add the
non-free-firmware section into sources.list for bookworm and later
releases.

-- 
regards Thomas

Bug#1031415: FAI fix

[Thomas Lange]

In FAI, we cannot easily determine which mke2fs or grub version will
be used in the target system since we support deb and rpm based and
other linux distributions.

As I did with the older issues of mke2fs (metadata_csum) I will add a
comment, so the user can decide if he needs to add the option
-O ^metadata_csum_seed

In FAI this has to be done in the config examples (which goes into the
package fai-doc), not in the code itself. I therefore decrease the
severity to normal.

-- 
regards Thomas

Bug#1029185: use alien

[Thomas Lange]

Hi,

I never build an rpm package, so I'm maybe the wrong person for doing
this. Plese try if the Debian package alien can help you building a
rpm packages that works on rpm systems.

   alien - Convert or install an alien binary package


$ fakeroot alien --to-rpm fai-client_6.0_all.deb
Warning: Skipping conversion of scripts in package fai-client: postrm
Warning: Use the --scripts parameter to include the scripts.
fai-client-6.0-2.noarch.rpm generated

-- 
regards Thomas

Bug#1000368: how to implement this using a hook

[Thomas Lange]

Hi Reiner,

here's an idea how you can implement it using a hook.
The hook is called for e.g. hooks/partition.DEFAULT
In this hook you concatenate the disk_config files for all classes
that match and write the output to disk_config/$HOSTNAME
Then setup-storage will use  this file. IIRC the FAI config space is
mounted read-only, but writeable using a ramdisk. So you won't modify
things on the server.

Here's some bash code:
for c in $classes; do
  if [ -f $FAI/disk_config/$c ]; then
cat $FAI/disk_config/$c >> $FAI/disk_config/$HOSTNAME
done

--
best regards Thomas

Bug#1029747: Empty "Categorize using"

[Thomas Lange]

> On Fri, 27 Jan 2023 09:18:20 +0100, Holger Wansing  
> said:

> This has been there for 14 years now. Wow!
> It was added in
> 
https://salsa.debian.org/webmaster-team/webwml/-/commit/4d417d00ddf442a347b1262e718f32a3a08d8e5d

> So I assume we can just remove it.
Yep, go ahead and remove it.

-- 
regards Thomas

Bug#1028303: dracut-core: missing Breaks+Replaces: dracut-live (<< 059-2)

[Thomas Lange]

Hi Andreas,

should I replace the old Breaks+Replaces with the new ones, or add them?
These are the current for dracut-core:
Breaks: dracut-network (<< 057+157-2)
Replaces: dracut-network (<< 057+157-2)

-- 
regards Thomas

Bug#1017039: more infos

[Thomas Lange]

I did some more tests.

modules.d/90overlayfs installs the script that mounts the overlay
using this line:

inst_hook mount 01 "$moddir/mount-overlayfs.sh"

But the mount hooks of dracut are not executed at all.
Here's the part of init.log:


/init@229(): getarg rd.break=mount -d rdbreak=mount
/lib/dracut-lib.sh@153(getarg): debug_off
/lib/dracut-lib.sh@23(debug_off): set +x
/lib/dracut-lib.sh@216(getarg): return 1
/init@232(): _i_mount=0
/init@233(): :
/init@234(): ismounted /sysroot
/lib/dracut-lib.sh@525(ismounted): findmnt /sysroot
/init@235(): usable_root /sysroot
/lib/dracut-lib.sh@736(usable_root): local _i
/lib/dracut-lib.sh@738(usable_root): '[' -d /sysroot ']'
/lib/dracut-lib.sh@740(usable_root): for _i in "$1"/usr/lib*/ld-*.so 
"$1"/lib*/ld-*.so
/lib/dracut-lib.sh@741(usable_root): '[' -e '/sysroot/usr/lib*/ld-*.so' ']'
/lib/dracut-lib.sh@740(usable_root): for _i in "$1"/usr/lib*/ld-*.so 
"$1"/lib*/ld-*.so
/lib/dracut-lib.sh@741(usable_root): '[' -e '/sysroot/lib*/ld-*.so' ']'
/lib/dracut-lib.sh@744(usable_root): for _i in proc sys dev
/lib/dracut-lib.sh@745(usable_root): '[' -e /sysroot/proc ']'
/lib/dracut-lib.sh@744(usable_root): for _i in proc sys dev
/lib/dracut-lib.sh@745(usable_root): '[' -e /sysroot/sys ']'
/lib/dracut-lib.sh@744(usable_root): for _i in proc sys dev
/lib/dracut-lib.sh@745(usable_root): '[' -e /sysroot/dev ']'
/lib/dracut-lib.sh@748(usable_root): return 0
/init@235(): break


Because of the break this line is not executed in init.sh and the
overlay mount script is not executed.

line 238 for f in "$hookdir"/mount/*.sh; do

-- 
regards Thomas

Bug#1025951: comment

[Thomas Lange]

You are right, it makes no sense to keep this patch. My justification
for the patch was wrong. I will fix that.
-- 
viele Grüße Thomas

Bug#1017039: 90overlay-root vs 90dmsquash-live

[Thomas Lange]

> On Tue, 20 Dec 2022 19:50:16 -0500, Laszlo  said:

> Can you please make sure that overlay-root is not included in initramfs.
Yes, it's included.
> Do you have rd.live.overlay.overlayfs=1 set on the command line ?
Yes.

I will do the rest soon.
> Can you please list the output of the dracut command when initramfs is
> generated ?
> Can you please perhaps post the full log ?

> Perhaps best to open a bug in
> https://github.com/dracutdevs/dracut/issues for this ?


-- 
viele Grüße Thomas

Bug#1017039: 90overlay-root vs 90dmsquash-live

[Thomas Lange]

> On Mon, 12 Dec 2022 17:33:18 -0500, Laszlo  said:

> 
https://github.com/dracutdevs/dracut/blob/master/test/TEST-20-NFS/test.sh#L178
I cannot reproduce that this works.
It always mounts the nfsroot read-only, but no writeable overlayfs in
tmpfs is available. I didn't found the problem yet, but I can't see
that the overlayfs module is executed (rd.debug enabled and grep in 
/run/initramfs/init.log)

-- 
regards thomas

Bug#1025951: systemd-users patch

[Thomas Lange]

I guess we still need thee patch, because I gues dracut in a non
systemd environment.

-- 
viele Grüße Thomas

Bug#964761: Creating btrfs subvolumes with setup-storage fails in certain conditions

[Thomas Lange]

Thanks a lot for the patch. I hoped it would be that easy. I've
already applied the patch.
https://github.com/faiproject/fai/commit/8bf5ef41461edf9c64fb1e8095fd12d4789cc6f4

The upcoming release FAI 6.0 will include it.

-- 
regards Thomas

Bug#1013904: info

[Thomas Lange]

The package list was too long because the repoquery command changed.
The new rinse version 4.x will say:

dnf repoquery --requires --resolve --recursive --installed ..

--installed changed a lot when generating the package list.

-- 
regards Thomas

Bug#1013904: package lists

[Thomas Lange]

Hi Benjamin,

I'm working on a newer rinse version and started to look into your
patches.

I wonder why for the fedora releases the package list includes things
like

fedora-release-identity-cinnamon
fedora-release-container

I see that they are small, but I'm not sure if they have dependencies
on other packages. BTW, how big are the chroot enviroments for the new
releases?


For fedora 36 I really wonder why these packages were added:

graphite2
bubblewrap
cairo
harfbuzz
pixman
flatpak
.
.
.

These don't fit in a minimal chroot IMO. So what is the reason to add
thoses packages?


-- 
regards Thomas

Bug#1024435: dracut-core: missing Breaks+Replaces: dracut-network (<< 057+157-2)

[Thomas Lange]

> On Sat, 19 Nov 2022 13:29:52 +0100, Andreas Beckmann  
> said:

>> From the attached log (scroll to the bottom...):

>   Preparing to unpack .../dracut-core_057+157-2_amd64.deb ...
>   Unpacking dracut-core (057+157-2) over (056-3) ...
>   dpkg: error processing archive 
/var/cache/apt/archives/dracut-core_057+157-2_amd64.deb (--unpack):
>trying to overwrite 
'/usr/lib/dracut/modules.d/95virtfs/module-setup.sh', which is also in package 
dracut-network 056-3
>   dpkg-deb: error: paste subprocess was killed by signal (Broken pipe)
>   Errors were encountered while processing:
>/var/cache/apt/archives/dracut-core_057+157-2_amd64.deb

I can understand the problem (and the fix to add the breaks relation),
but I wonder why I should define a replaces to dracut-network?
dracut-core does not replace dracut-network. It only breaks
dracut-network (<< 057+157-2). Can you please tell me what correct way
is to fix it.

-- 
regards Thomas

Bug#902061: possible solution

[Thomas Lange]

I found an example how it may be possible to use lvm names:
NEWS.md says:

$ dracut --print-cmdline
rd.lvm.lv=debian/root rd.lvm.lv=debian/usr root=/dev/mapper/debian-root


The rd.lvm.lv parameters are only needed if you do not want to
activate all lv's. You may need to add rd.auto for autoassembly of all
lvm devices. See man dracut.cmdline(7) for details..

-- 
regards Thomas