Re: Daily Images Fail to Install - Debian

This is a discussion on Re: Daily Images Fail to Install - Debian ; * Gordon Farquharson [2007-11-24 00:16]: > /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot > initialize debconf template database > > Below is the boot log from an NSLU2 (arm/ixp4xx). > > Is this a known problem? FWIW, I just booted a daily image on ...

+ Reply to Thread
Results 1 to 9 of 9

Thread: Re: Daily Images Fail to Install

  1. Re: Daily Images Fail to Install

    * Gordon Farquharson [2007-11-24 00:16]:
    > /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot
    > initialize debconf template database
    >
    > Below is the boot log from an NSLU2 (arm/ixp4xx).
    >
    > Is this a known problem?


    FWIW, I just booted a daily image on Cobalt (mipsel) and didn't see
    this.
    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  2. Re: Daily Images Fail to Install

    * Martin Michlmayr [2007-11-24 11:09]:
    > > /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot
    > > initialize debconf template database
    > >
    > > Is this a known problem?

    > FWIW, I just booted a daily image on Cobalt (mipsel) and didn't see
    > this.


    I believe this is caused by a binutils bug on ARM. I just built an
    image on ARM myself and noticed that ld produces an assertation. Can
    you look through your logs? (grep for "assertion fail")

    I reported this upstream to http://sourceware.org/bugzilla/show_bug.cgi?id=5398
    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  3. Re: Daily Images Fail to Install

    * Martin Michlmayr [2007-11-24 20:07]:
    > * Martin Michlmayr [2007-11-24 11:09]:
    > > > /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot
    > > > initialize debconf template database
    > > >
    > > > Is this a known problem?

    > > FWIW, I just booted a daily image on Cobalt (mipsel) and didn't see
    > > this.

    >
    > I believe this is caused by a binutils bug on ARM. I just built an
    > image on ARM myself and noticed that ld produces an assertation. Can
    > you look through your logs? (grep for "assertion fail")


    Bad news. I just used a fix binutils to build a daily image and the
    same happens to me on NSLU2. Joey/Colin, any idea what might be going
    on?

    When d-i starts it only repeats the following message:

    x1205 0-006f: setting the system clock to 2007-11-28 16:47:16 (1196268436)
    Freeing init memory: 100K
    /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot initialize debconf template database
    /build/buildd/cdebconf-0.125/src/debconf.c:135 (main): Cannot initialize debconf template database
    ....

    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  4. Re: Daily Images Fail to Install

    * Martin Michlmayr [2007-11-28 18:09]:
    > Bad news. I just used a fix binutils to build a daily image and the
    > same happens to me on NSLU2. Joey/Colin, any idea what might be
    > going on?


    Okay, I broke the ARM images with my Netwinder addition.
    /lib/debian-installer-startup.d/S02netwinder-net runs archdetect
    before the debconf data is loaded (I guess in S20templates).
    Unfortunately, S02netwinder-net has to run before udev is started,
    otherwise the udev rule it creates won't become active.

    Can we move S20templates to the front, or is there a way to reload
    udev after netwinder-net created the udev rule? Or any other solution
    I cannot think of?
    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  5. Re: Daily Images Fail to Install

    On Wednesday 28 November 2007, Martin Michlmayr wrote:
    > Okay, I broke the ARM images with my Netwinder addition.
    > /lib/debian-installer-startup.d/S02netwinder-net runs archdetect
    > before the debconf data is loaded (I guess in S20templates).
    > Unfortunately, S02netwinder-net has to run before udev is started,
    > otherwise the udev rule it creates won't become active.


    To be honest, I don't see how archdetect requires or should require debconf
    templates. Are you sure archdetect is the cause?
    If it is, maybe that is something that could be fixed.

    > Can we move S20templates to the front,


    I don't think that is a very nice solution. There is something like a
    natural order of things and this would IMO go against that.

    > or is there a way to reload udev after netwinder-net created the udev
    > rule?


    No idea.

    > Or any other solution I cannot think of?


    Detect the subarch in a different way (by parsing /proc/cpuinfo directly)?

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHTkpugm/Kwh6ICoQRAkkZAJ9M2IGLAtHUktUlsuThlkWQK/CLbgCfTubb
    A27jEGIPOZPnAA0qGm1obpo=
    =n56h
    -----END PGP SIGNATURE-----


  6. Re: Daily Images Fail to Install

    * Frans Pop [2007-11-29 06:13]:
    > To be honest, I don't see how archdetect requires or should require
    > debconf templates. Are you sure archdetect is the cause?


    Actually, my diagnosis was wrong. For some reason I thought
    archdetect started using debconf a few months ago to cache the result
    in a debconf variable, but I just checked the source and I'm wrong.

    Anyway, I found the real cause: if the machine is not Netwinder, the
    script does an "exit 0" - but we're being sourced rather than run, so
    this obviously has to be "return 0". This also explains why I didn't
    see the problem when I tested on a Netwinder. I'll upload a fix.
    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  7. Re: Daily Images Fail to Install

    On Thursday 29 November 2007, Martin Michlmayr wrote:
    > Anyway, I found the real cause: if the machine is not Netwinder, the
    > script does an "exit 0" - but we're being sourced rather than run, so
    > this obviously has to be "return 0". This also explains why I didn't
    > see the problem when I tested on a Netwinder. I'll upload a fix.


    I doubt return 0 works better when sourcing.
    You'll probably have to bring the actual code inside the if statement (and
    reverse the condition).

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHTnYGgm/Kwh6ICoQRAgtjAKCsJjPbb/8wBmhdBgUAWQ+bO+OqlgCfddAI
    E7C3hdXCFLuUFXe/+3Fa20A=
    =Ro32
    -----END PGP SIGNATURE-----


  8. Re: Daily Images Fail to Install

    * Frans Pop [2007-11-29 09:19]:
    > > Anyway, I found the real cause: if the machine is not Netwinder, the
    > > script does an "exit 0" - but we're being sourced rather than run, so
    > > this obviously has to be "return 0". This also explains why I didn't
    > > see the problem when I tested on a Netwinder. I'll upload a fix.

    > I doubt return 0 works better when sourcing.


    Are you sure? It seems to work for me...

    > You'll probably have to bring the actual code inside the if
    > statement (and reverse the condition).


    .... but I can do that instead.
    --
    Martin Michlmayr
    http://www.cyrius.com/


    --
    To UNSUBSCRIBE, email to debian-boot-REQUEST@lists.debian.org
    with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

  9. Re: Daily Images Fail to Install

    On Thursday 29 November 2007, Martin Michlmayr wrote:
    > * Frans Pop [2007-11-29 09:19]:
    > > > Anyway, I found the real cause: if the machine is not Netwinder, the
    > > > script does an "exit 0" - but we're being sourced rather than run, so
    > > > this obviously has to be "return 0". This also explains why I didn't
    > > > see the problem when I tested on a Netwinder. I'll upload a fix.

    > >
    > > I doubt return 0 works better when sourcing.

    >
    > Are you sure? It seems to work for me...


    /me reads up on sourcing in bash manpage
    You could be right. I thought sourcing meant that anything in the file
    effectively became part of the current script, but it talks about "read and
    execute", so it could be more like calling a function.

    -----BEGIN PGP SIGNATURE-----
    Version: GnuPG v1.4.6 (GNU/Linux)

    iD8DBQBHToEtgm/Kwh6ICoQRAlEeAKCRhjhdA7Z2aaK2ShoYpiIGK3JOYgCfXVe8
    iEWDZkIWJADXhRW4Gh2/cKs=
    =iG6k
    -----END PGP SIGNATURE-----


+ Reply to Thread