Bug#965263: debian-installer: Script exected in preseed/late_command on dual CPU socket system sees only Single CPU socket

[Vasudev Kamath]

Package: debian-installer
Severity: normal

Dear Maintainer,

I recently moved some code to preseed/late_command script to do the CPU pinning 
for host machine.

Script snippet looks something like below

NODE_CPUS=2
for node in /sys/devices/system/node/node[0-9]; do
cat $node/cpu[0-9]*/topology/thread_siblings_list | sort 
-nu | sed 's/,/\n/' | head -n$NODE_CPUS
done

It captured a sibling core from only node0 of the system. So out of curiosity I 
observed the logs we captured
on rsyslog server and found what I see below

Jul 14 09:45:18 10.33.97.110 in-target: ++ cat 
/sys/devices/system/node/node0/cpu0/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu10/topology/thread_siblings_list 
/sys/devices/system/no
de/node0/cpu11/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu12/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu13/topology/thread_siblings_list 
/sys/devices/system/nod
e/node0/cpu14/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu15/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu16/topology/thread_siblings_list 
/sys/devices/system/node
/node0/cpu17/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu18/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu19/topology/thread_siblings_list 
/sys/devices/system/node/
node0/cpu1/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu20/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu21/topology/thread_siblings_list 
/sys/devices/system/node/no
de0/cpu22/topology/thr
Jul 14 09:45:18 10.33.97.110 in-target: 
system/node/node0/cpu23/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu24/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu25/topo
logy/thread_siblings_list 
/sys/devices/system/node/node0/cpu26/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu27/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu28/topol
ogy/thread_siblings_list 
/sys/devices/system/node/node0/cpu29/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu2/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu30/topolog
y/thread_siblings_list 
/sys/devices/system/node/node0/cpu31/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu32/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu33/topology
/thread_siblings_list 
/sys/devices/system/node/node0/cpu34/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu35/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu36/topology/
thread_siblings_list /
Jul 14 09:45:18 10.33.97.110 in-target: pu37/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu38/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu39/topology/thread_sibling
s_list /sys/devices/system/node/node0/cpu3/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu40/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu41/topology/thread_siblings_
list /sys/devices/system/node/node0/cpu42/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu43/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu44/topology/thread_siblings_l
ist /sys/devices/system/node/node0/cpu45/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu46/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu47/topology/thread_siblings_li
st /sys/devices/system/node/node0/cpu48/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu49/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu4/topology/thread_siblings_list
 /sys/devices/system/n
Jul 14 09:45:18 10.33.97.110 in-target: _siblings_list 
/sys/devices/system/node/node0/cpu51/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu52/topology/thread_siblings_list /sys/devices/
system/node/node0/cpu53/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu54/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu55/topology/thread_siblings_list 
/sys/devices/s
ystem/node/node0/cpu56/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu57/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu58/topology/thread_siblings_list 
/sys/devices/sy
stem/node/node0/cpu59/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu5/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu60/topology/thread_siblings_list 
/sys/devices/syst
em/node/node0/cpu61/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu62/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu63/topology/thread_siblings_list 
/sys/devices/syste
m/node/node0/cpu64/top
Jul 14 09:45:18 10.33.97.110 in-target: 
/devices/system/node/node0/cpu65/topology/thread_siblings_list 
/sys/devices/system/node/node0/cpu66/topology/thread_siblings_list 
/sys/devices/system/nod

Bug#965263: debian-installer: Script exected in preseed/late_command on dual CPU socket system sees only Single CPU socket

[Geert Stappers]

On Sat, Jul 18, 2020 at 06:19:55PM +0530, Vasudev Kamath wrote:
> 
> Please let me know if you need more information.
> 

Both kernel versions

Re: Installation Issue

[Samuel Thibault]

Hello,

Vlad Dragomir, le mar. 14 juil. 2020 21:11:59 +0200, a ecrit:
> First I'd like to apologise for asking this here. It's not strictly related 
> to accessibility, but since I'm using the accessible Speakup driven 
> installer, I thought it might be easier for members of this list to guide me 
> if they are able to.

Well, possibly when it's related to accessibility, but here it's
really not, and thus you won't get expertise on the matter from
debian-accessibility@. I have thus Cc-ed debian-boot@, since that's
where expertise on the matter is. Normally, debian-boot should be able
to help non-sighted people too.

> I tried to install it on an unused laptop I'm having, using the iso 
> downloaded from this torrent:
> https://cdimage.debian.org/cdimage/unofficial/non-free/cd-including-firmware/10.4.0+nonfree/amd64/bt-dvd/firmware-10.4.0-amd64-DVD-1.iso.torrent
> 
> It started wonderfully, speech came  up immediately and the installation went 
> smoothly until it tried to connect to the Internet. It asks for additional 
> non-free firmware, unfortunately it's an Intel wireless card. It gives a list 
> of files it wants in order to configure the network, which is I suppose a 
> good start. Still,


> 3.This installation iso pretends it contains non-free firmware already, 
> so normally it shouldn’t ask for those files. Am I missing something?

You didn't miss anything, that should be it.

> Therefore, I had to cancel the installation, at least for the time being. It 
> would be rather useless to have a system without Internet access. 
> 
> Here is the list of files the installer asks for:
> iwlwifi-8000C-36.ucode
> iwlwifi-8000C-35.ucode
> iwlwifi-8000C-34.ucode
[... down to...]
> iwlwifi-8000C-22.ucode

It's surprising: they should indeed be already there, since they're in
the firmware-iwlwifi non-free package, even in version 20190114-2 that the
10.4.0 image ships.

I don't know more about the potential tricks with iwlwifi, thus leaving
up to debian-boot.

Samuel

Bug#690210: strange sources.list when using debootstrap with option --extra-suite

[jhcha54008]

Hi again,

It seems that components are duplicated when using option --extra-suite in 
debootstrap.

I got the following /etc/apt/sources.list (note the 'main' component repeated 
twice or four times)

   --8<--/etc/apt/sources.list---8<--
deb http://deb.debian.org/debian stable main main
deb http://deb.debian.org/debian buster-proposed-updates main main main main
   --8<--/etc/apt/sources.list---8<--

as a result of the invocation of debootstrap (version 1.0.123) :

# debootstrap --variant=minbase --extra-suites=buster-proposed-updates stable \
  my_stable_chroot_with_updates

The partial fix for bug #690210 seems to be the culprit. I recover the expected 
sources.list with the patch below.

Thank you for your ongoing work on debootstrap !

Regards,
JH Chatenet 

diff -Naur a/usr/share/debootstrap/functions b/usr/share/debootstrap/functions
--- a/usr/share/debootstrap/functions   2020-03-14 02:07:20.0 +0100
+++ b/usr/share/debootstrap/functions   2020-07-15 18:49:15.214419623 +0200
@@ -547,6 +547,9 @@
mv "$reldest" "$reldest.malformed"
error 1 INVALIDREL "Invalid Release file, no valid components"
fi
+
+   # Suppress duplicated components (this may occur if EXTRA_SUITES are 
selected)
+   COMPONENTS=$(echo "$COMPONENTS" | tr ' ' '\n' | sort | uniq | tr '\n' ' 
')
 }
 
 CODENAME=""
@@ -1110,6 +1113,7 @@
for m in "$@"; do
for s in $SUITE $EXTRA_SUITES; do
local cs c path pkgdest
+   cs=""
for c in ${COMPONENTS:-$USE_COMPONENTS}; do
path="dists/$s/$c/binary-$ARCH/Packages"
pkgdest="$TARGET/$($DLDEST pkg "$s" "$c" 
"$ARCH" "$m" "$path")"

Re: stretch EOL point release (9.13) and 10.5 planning

[Adam D. Barratt]

On Sun, 2020-07-12 at 15:35 +0100, Adam D. Barratt wrote:
> Hi Steve,
> 
> On Sun, 2020-07-12 at 12:46 +0100, Steve McIntyre wrote:
> > Argh, massive apologies...
> > 
> > On Thu, Jun 25, 2020 at 10:38:14AM +0200, Laura Arjona Reina wrote:
> > > El 15/6/20 a las 18:44, Adam D. Barratt escribió:
> > > > - July 18/19
> > 
> > Massive apologies for dropping a spanner in the works, but
> > something major has come up. I won't be able to do *all* of that
> > weekend after all. As Stretch EOL is already a thing, can I suggest
> > that we keep that to plan and push back the Buster 10.5 release a
> > little?
> > 
> > Sorry. :-/
> 
> Thanks for letting us know. :-(
> 
> I'll drop a note to the lists, and we can look at getting a new date
> organised.

Now that stretch EoL is (more or less) out of the way, it would be good
to get 10.5 done as soon as we sensibly can, so as not to slip too far
off schedule.

Next weekend is probably a little too soon - I'd at least like to not
jump straight back into freezing - but how about one of:

- August 1st/2nd
- August 8th/9th

Regards,

Adam

Re: stretch EOL point release (9.13) and 10.5 planning

[Cyril Brulebois]

Adam D. Barratt  (2020-07-18):
> Now that stretch EoL is (more or less) out of the way, it would be good
> to get 10.5 done as soon as we sensibly can, so as not to slip too far
> off schedule.
> 
> Next weekend is probably a little too soon - I'd at least like to not
> jump straight back into freezing - but how about one of:
> 
> - August 1st/2nd
> - August 8th/9th

Both should work equally for me regarding the installer.


Cheers,
-- 
Cyril Brulebois (k...@debian.org)
D-I release manager -- Release team member -- Freelance Consultant


signature.asc
Description: PGP signature

Re: stretch EOL point release (9.13) and 10.5 planning

[Steve McIntyre]

On Sat, Jul 18, 2020 at 10:12:41PM +0100, Adam Barratt wrote:
>On Sun, 2020-07-12 at 15:35 +0100, Adam D. Barratt wrote:
>> Hi Steve,
>> 
>> On Sun, 2020-07-12 at 12:46 +0100, Steve McIntyre wrote:
>> > Argh, massive apologies...
>> > 
>> > On Thu, Jun 25, 2020 at 10:38:14AM +0200, Laura Arjona Reina wrote:
>> > > El 15/6/20 a las 18:44, Adam D. Barratt escribió:
>> > > > - July 18/19
>> > 
>> > Massive apologies for dropping a spanner in the works, but
>> > something major has come up. I won't be able to do *all* of that
>> > weekend after all. As Stretch EOL is already a thing, can I suggest
>> > that we keep that to plan and push back the Buster 10.5 release a
>> > little?
>> > 
>> > Sorry. :-/
>> 
>> Thanks for letting us know. :-(
>> 
>> I'll drop a note to the lists, and we can look at getting a new date
>> organised.
>
>Now that stretch EoL is (more or less) out of the way, it would be good
>to get 10.5 done as soon as we sensibly can, so as not to slip too far
>off schedule.
>
>Next weekend is probably a little too soon - I'd at least like to not
>jump straight back into freezing - but how about one of:
>
>- August 1st/2nd
>- August 8th/9th

Either is possible for me, with a preference for the first. Let's not
delay too long if possible.

Cheers,

Steve

-- 
Steve McIntyre, Cambridge, UK.st...@einval.com
  Armed with "Valor": "Centurion" represents quality of Discipline,
  Honor, Integrity and Loyalty. Now you don't have to be a Caesar to
  concord the digital world while feeling safe and proud.

Re: stretch EOL point release (9.13) and 10.5 planning

[Andy Simpkins]

On 18 July 2020 22:39:05 BST, Steve McIntyre  wrote:
>On Sat, Jul 18, 2020 at 10:12:41PM +0100, Adam Barratt wrote:
>>On Sun, 2020-07-12 at 15:35 +0100, Adam D. Barratt wrote:
>>> Hi Steve,
>>> 
>>> On Sun, 2020-07-12 at 12:46 +0100, Steve McIntyre wrote:
>>> > Argh, massive apologies...
>>> > 
>>> > On Thu, Jun 25, 2020 at 10:38:14AM +0200, Laura Arjona Reina
>wrote:
>>> > > El 15/6/20 a las 18:44, Adam D. Barratt escribió:
>>> > > > - July 18/19
>>> > 
>>> > Massive apologies for dropping a spanner in the works, but
>>> > something major has come up. I won't be able to do *all* of that
>>> > weekend after all. As Stretch EOL is already a thing, can I
>suggest
>>> > that we keep that to plan and push back the Buster 10.5 release a
>>> > little?
>>> > 
>>> > Sorry. :-/
>>> 
>>> Thanks for letting us know. :-(
>>> 
>>> I'll drop a note to the lists, and we can look at getting a new date
>>> organised.
>>
>>Now that stretch EoL is (more or less) out of the way, it would be
>good
>>to get 10.5 done as soon as we sensibly can, so as not to slip too far
>>off schedule.
>>
>>Next weekend is probably a little too soon - I'd at least like to not
>>jump straight back into freezing - but how about one of:
>>
>>- August 1st/2nd
>>- August 8th/9th
>
>Either is possible for me, with a preference for the first. Let's not
>delay too long if possible.
>
>Cheers,
>
>Steve
>
>-- 
>Steve McIntyre, Cambridge, UK.   
>st...@einval.com
>  Armed with "Valor": "Centurion" represents quality of Discipline,
>  Honor, Integrity and Loyalty. Now you don't have to be a Caesar to
>  concord the digital world while feeling safe and proud.

-- 
I am good for either weekend.
Given a preferance i would prefer the 1st.  

Cheers
/Andy
Sent from my Android device with K-9 Mail. Please excuse my brevity.

RE: Installation Issue

[Vlad Dragomir]

Thank you very much Samuel, I really didn’t know about Debian Boot. I asked on 
the Debian User list instead, and a few very nice people kindly sent me 
suggestions which finally allowed me to finish the installation.

 

I’d like to apologise once more for not knowing about the impressive number of 
mailing lists before posting here. I’ll only ask about accessibility here in 
the future, now I’m starting to understand how everything is being organised.

 

Best regards and thank you one last time.

Bug#965263: debian-installer: Script exected in preseed/late_command on dual CPU socket system sees only Single CPU socket

[Vasudev Kamath]

Geert Stappers  writes:

> On Sat, Jul 18, 2020 at 06:19:55PM +0530, Vasudev Kamath wrote:
>> 
>> Please let me know if you need more information.
>> 
>
> Both kernel versions

We mirror Debian repository on daily basis and found this bug with
latest installer released with Stretch 9.12 and Buster 10.4.

Buster: linux-image-4.19.0-9-amd64 (4.19.118-2+deb10u1)
Stretch: linux-image-4.9.0-12-amd64 (4.9.210-1+deb9u1)

Cheers,
Vasudev