I'm going to try to dig into this and see if I can fix it. But if anyone
knows off the top of their head where a likely place to start looking would
be, please let me know. Otherwise I'll start with the bootstrap scripts.
- Jason
ped working.
I've attached the makefile and a log of a failed build.
- Jason Heeris
Makefile
Description: Binary data
[2017-01-19 11:57:22] lb bootstrap
P: Setting up cleanup function
[2017-01-19 11:57:22] lb bootstrap_cache restore
[2017-01-19 11:57:22] lb bootstrap_debootstrap
P: Begin
Okay, after clearing caches, cleaning up env vars, etc. I'm able to
reliably reproduce the problem.
I've attached two makefiles, one that works (in that, it builds an image, I
don't care about running the image yet). The other fails to build with the
aforementioned error. In the last version of li
On 16 January 2017 at 14:35, Jason Heeris wrote:
> On 16 January 2017 at 14:20, Michael . wrote:
> > If you are building a Raspbian image your initial config should work
> I will try again and see if removing the firmware section makes a
difference with a Raspbian build.
Turns out
On 16 January 2017 at 14:20, Michael . wrote:
> If you are building a Raspbian image your initial config should work
But that's the problem, it doesn't. See:
https://gist.github.com/detly/109b8201e7c81417af3a8f455f415a6d
I will try again and see if removing the firmware section makes a
differenc
On 16 January 2017 at 14:08, Jason Heeris wrote:
> Here's a command to try building from a Raspbian image.
s/ from//
- Jason
Here's a command to try building from a Raspbian image.
env
DEBOOTSTRAP_OPTIONS="--keyring=/usr/share/keyrings/raspbian-archive-keyring.gpg"
lb config \
--archive-areas 'main firmware non-free' \
--distribution jessie \
--chroot-filesystem squashfs \
--architectures armhf \
--b
On 16 January 2017 at 10:21, Ben Armstrong
wrote:
>
> I can't see how --archive-areas 'main firmware' would've ever worked,
since 'firmware' is not valid.
Hmm. It was left over from my original project which was a Raspbian live
build, which suddenly started failing in the same way. I tried to red
Forgot to mention: this looks a lot like bugs #813232 and #571136.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=571136
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=813232
- Jason
his broke somewhere between version 1:20151215 and 1:20161216.
Full boot log is here:
https://gist.github.com/detly/a0752dfafcdeb1eb8f2859c41a63fa32
Cheers,
Jason Heeris
On 14 February 2016 at 13:01, Jason Heeris wrote:
> I am at a loss trying to explain this auto/config word splitting
> behaviour.
Okay, it turns out this is a well known... puzzle in shell scripting.
Apologies for the noise, but hey, I learnt a lot today.
— Jason
On 14 February 2016 at 13:30, Jason Heeris wrote:
> Well, I manually typed the test script into another file, and ran lb
> config, and it worked.
Haha, no it didn't. I forgot to chmod +x it, so lb config ignored it.
The problem remains. Any advice is appreciated.
Cheers,
Jason
On 14 February 2016 at 13:01, Jason Heeris wrote:
> I am at a loss trying to explain this auto/config word splitting
> behaviour.
Well, I manually typed the test script into another file, and ran lb
config, and it worked. I replaced the old file with my new one, and
git told me there was
I am at a loss trying to explain this auto/config word splitting
behaviour. My auto/config contains only:
--
#!/bin/sh
set -e
GENERAL_BUILD_OPTIONS="--gzip-options '-9 --rsyncable' --security false"
lb config noauto ${GENERAL_BUILD_OPTIONS} "${@}"
--
(This is not my actual project, it's just a
filesystems...
P: Saving caches...
/usr/bin/env: apt-get: No such file or directory
The full log is attached.
Cheers,
Jason Heeris
PS. Sorry for (a) not using reportbug, but I'm not at that machine right
now;
and (b) the multiple postings.
build_log
Description: Binary data
ve or -boot issue. (Also,
sorry for breaking threading, but I wasn't subscribed to -boot.)
Cheers,
Jason Heeris
use a
different keyring? (I don't really want to tell it not to verify anything
at all...)
I realise Raspbian questions might not be strictly on topic, and I
apologise if that's the case, but I feel like it's a simple question of how
debootstrap is verifying its downloads, and this seems like a good place to
ask about that.
Cheers,
Jason Heeris
I've been trying to create a Debian live image that will boot via UEFI, but
haven't been able to do so. My live-build host VM runs Debian 9 (stretch),
with live build 4.0.3-1. I noticed that bug #731709 is open [1], but no
mention is made of an upcoming fix.
The config structure I'm using is here:
On 5 May 2015 at 02:43, Daniel Baumann <
daniel.baum...@progress-technologies.net> wrote:
> if you want to build wheezy or jessie images, stick with 4.x; for
> stretch/sid use 5.x.
Do either of 4.x/5.x support UEFI booting? There's an open bug about it
(#731709), but the Debian installer image see
detly/dart/tree/jessie-only
I have also tried a bootappend-live line of nothing but:
boot=live components autologin
Even with this line, I'm prompted for credentials. How can I get autologin
to work?
— Jason Heeris
On 4 April 2012 17:56, Jason Heeris wrote:
> Will do! My search continues. I might (a) try some packages from
> wheezy, and (b) ask on the user list too.
So I want to try to pull in udisks and all dependencies from sid. I
have config/archives/sid.list.chroot containing a Debian sid lin
On 4 April 2012 17:21, Daniel Baumann
wrote:
>> T0:23:respawn:/bin/login -f user /dev/ttyS0 2>&1
>
> we probably should add that into live-config itself, will look into that
> somewhen in the next days.
There's already a bug report for it, #602788
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug
On 3 April 2012 09:49, Jason Heeris wrote:
> no
> no
> yes
>
> I'll try changing the "allow_inactive" key to yes, but this seems less
> than ideal
It doesn't work anyway, I'm still not authorised to mount disks using
udisks over a serial console.
I'm building a Debian Squeeze system with live-builder 3.0~a45 (from
Ubuntu 11.10). I'm running into a problem where I can't use udisk to
mount devices using a serial console on the live system (perhaps due
to problems with consolekit or policykit).
I configure a serial console using the boot line
I've dropped the idea of using persistence for this, and just done
some hackery with udisks and bash scripting. It seems to work fine.
I'm still not sure what I was doing wrong with the
persistent-read-only option, but I suppose that's only interesting if
it really is a bug and not just my mistake
On 22 March 2012 13:04, Jason Heeris wrote:
> However I still can't seem to get the
> "persistent-read-only" option to work. It doesn't seem to do anything
Okay, it seems that if I use the two options "persistent" and
"persistent-read-only&quo
On 21 March 2012 13:16, Jason Heeris wrote:
> The most recent version of live-boot has an
> option "persistent-read-only" which would get around this, that this
> is not available in the version installed on a Squeeze system
> (2.0.15-1).
Okay, so I have the unstable versio
I'm using live-build 3.0~a45-1 under Ubuntu 11.10 to create a live
Squeeze system on a USB drive. I have a file (~500MB) that I want on
the final image, so at the moment I just put it in
config/includes.chroot/opt/stuff and it gets bundled in when I build
the image. This was fine when the image cha
On 7 March 2012 10:10, Diederik de Haas wrote:
> I don't have experience with squid(-deb-proxy), so I don't know how to
> configure that.
> I use apt-cacher-ng and the command I used was:
> lh config --mirror-bootstrap
> "http://192.168.1.5:3142/ftp.nl.debian.org/debian/"; --mirror-binary
> "htt
On 7 March 2012 10:10, Diederik de Haas wrote:
> On Wednesday 07 March 2012 02:44:28 Jason Heeris wrote:
> I don't have experience with squid(-deb-proxy), so I don't know how to
> configure that.
> I use apt-cacher-ng and the command I used was:
Aha! Searching for "a
On 7 March 2012 09:24, Diederik de Haas wrote:
> I realize that during one of the build stages you get a warning that if that
> stage fails it's likely
> caused by using a (deb) proxy, but I've had good experience with using a
> (deb) proxy, so it may be
> worth a shot.
> Nice side effect is tha
I'm trying to build a live Wheezy system, and since I live in
Australia I'd like to use an Australian mirror. Unfortunately, every
time I've tried to use one over the last ten days, my build has
failed. Mirrors I've tried include the official ftp.au.debian.org,
ftp.netspace.net.au, mirror.aarnet.ed
Daniel's suggestion did the trick, but I don't know where it's documented :)
Is this the same as "ip=frommedia"?
— Jason
--
To UNSUBSCRIBE, email to debian-live-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive:
http://lists.debian.o
I have a Wheezy live system, built with live-build 3.0~a45-1 on an
Ubuntu 11.10 system. I've enabled persistence to a COW live-rw
partition, and I'm sure that it's working because when I write a test
file to /etc, it's there on subsequent boots.
The problem is: any changes I make to /etc/network/i
I've attached a hacky little sed script that Works For Me (goes in
config/includes.chroot/lib/live/config) — it requires manual tweaking for
different device names, and is certainly no replacement for the original
live-getty script. But it might help others who just need to get something
up and run
This is affecting me too. FYI, you can test serial support for a live image
with something like:
$ qemu -hda binary.img -serial stdio
Serial console then appears on the command line.
I've poked around to try to fix it (or work around it), but haven't really
gotten anywhere. Any tips on where
On 22 February 2012 10:24, Jason Heeris wrote:
>
> Okay, answering part of my own question here, but I found bug #442289[1]
> and saw that the solution is to add "live-getty" to the linux command line:
>
Bah, I spoke too soon. No error messages, but no serial console at all.
— Jason
On 21 February 2012 14:28, Jason Heeris wrote:
> But instead of a console on the serial line, I see:
>
> /bin/sh: cannot open /dev/tty: No such device or address
>
Okay, answering part of my own question here, but I found bug #442289[1]
and saw that the solution is to add &qu
On 21 February 2012 14:28, Jason Heeris wrote:
> 2. I copied the chroot "/etc/inittab" to
> "config/includes.chroot/etc/inittab" and uncommented this to spawn a serial
> console (note the change of speed, too)
>
> T0:2345:respawn:/sbin/getty -L ttyS0 11
I am building a live image using live-build 3.0~a45-1 on Ubuntu 11.10. The
target image is Debian Squeeze using linux-image-2.6.
I want a console on both the screen and the serial line, so I added the
following line to my auto/config:
LB_BOOTAPPEND_LIVE="quiet ipv6.disable=1 console=tty0
cons
On 22 November 2010 17:38, Daniel Baumann wrote:
> provide the complete build log, make sure it's not a local error, etc
> pp. basically, do the research as outlined in
> http://live.debian.net/manual/en/html/bugs.html.
Yes, my apologies for that. I was already partway through this
process, but t
On 19 November 2010 15:07, Jason Heeris wrote:
> On 18 November 2010 12:10, Tzafrir Cohen wrote:
>> Maybe explicitly install ssmtp (or some other minimal package)?
>
> By the way, I gave this a shot too — I created
> "chroot_local-packageslists/extras.list" con
On 18 November 2010 12:10, Tzafrir Cohen wrote:
> Maybe explicitly install ssmtp (or some other minimal package)?
By the way, I gave this a shot too — I created
"chroot_local-packageslists/extras.list" containing:
anacron
ssmtp
...but exim4-daemon-light still gets installed, and no sign of ssmt
On 19 November 2010 07:07, Jason Heeris wrote:
> The pinning idea is probably the best approach for me, if it works, so
> thanks Ben :)
Ah, I don't think apt-pinning works on virtual packages. Maybe a
chroot hook script then.
— Jason
--
To UNSUBSCRIBE, email to debia
The pinning idea is probably the best approach for me, if it works, so
thanks Ben :)
If anyone is interested, the reason that the "manually install all
recommends" approach is bad for me is that (a) the number of
recommends I don't want is less than a tenth of the number I'm fine
with keeping, and
On 18 November 2010 17:37, Daniel Baumann wrote:
> the right way imho is to go is to disable automatic installation of
> recommends and picking explicitly what you want. the 'i want all
> recommends but not $this and $that one' doesn't make any sense to me.
I was afraid of that, but okay. It chan
On 18 November 2010 12:10, Tzafrir Cohen wrote:
> Maybe explicitly install ssmtp (or some other minimal package)?
> Specifically, ssmtp does not start any daemon. But it Provides
> mail-transport-agent.
But I don't want an MTA *at all* — it's always the first thing I
remove from a new system, sin
I'm creating a fairly simple rescue USB drive, tailored towards a
specific low-end machine I'm working on. I'd like to be able to
prevent any MTA from being installed (amongst others). Since it's only
a recommended package this should be possible without breaking things,
so what is the simplest way
48 matches
Mail list logo