Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default - Debian

This is a discussion on Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default - Debian ; Hello, Christian Perrier forwarded me the question, as a longtime Solaris and OpenSolaris user, so let me try to shed more light on this. The executive summary is, yes, by default, on *x86* systems, both Solaris and OpenSolaris considers the ...

+ Reply to Thread
Results 1 to 3 of 3

Thread: Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default

  1. Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default

    Hello,

    Christian Perrier forwarded me the question, as a longtime Solaris and
    OpenSolaris user, so let me try to shed more light on this.

    The executive summary is, yes, by default, on *x86* systems, both
    Solaris and OpenSolaris considers the system clock to be the local time.

    The longer explanation:
    This is a very old convention, that probably dates back from when Sun
    acquired Interactive Unix for i386/i486 systems, to merge it into
    Solaris and make it Solaris x86.
    The manual dates back at least from Solaris 2.4 in 1994:
    http://docs.sun.com/app/docs/doc/801...4j?l=en&a=view

    I'm not sure how the Debian installer works on SPARC, and if it allows
    multiboot there, but this feature is x86 only. Solaris SPARC, and most
    probably the upcoming OpenSolaris SPARC use UTC, as shown in rtc(1M):

    DESCRIPTION
    On x86 systems, the rtc command reconciles the difference in
    the way that time is established between UNIX and MS-DOS
    systems. UNIX systems utilize Greenwich Mean Time (GMT),
    while MS-DOS systems utilize local time.

    The rtc command is run in the crontab at 02:01 every day. It's a no-op
    on SPARC. On x86, when it's the day for changing the time, it adjusts
    the system time accordingly so it stays on local time, and keeps track
    there of the time difference between system time and UTC.

    Thus in CET, I have:
    zone_info=Europe/Paris
    zone_lag=-3600

    And in CEST, it's:
    zone_info=Europe/Paris
    zone_lag=-7200


    Yes, as has been pointed out, it's definitely an issue if the system
    isn't up when the time changes. It will only be updated on the next time
    the command is run, which can take a while depending on the conditions
    the system is run.

    Also, the system can be set to UTC manually, but it's not very common
    practice, even on Sun x86 hardware.

    FWIW, there's an open bug about this, but I don't think it'll go
    anywhere fast:
    http://bugs.opensolaris.org/view_bug.do?bug_id=6395979

    And yes, it's clunky, and has been the cause of at least one bad issue:
    http://bugs.opensolaris.org/view_bug.do?bug_id=6247281

    I wouldn't point a finger at anybody for the blame. When it was
    introduced, MS-DOS compatibility was important. It still is, and both
    the PC hardware manufacturers and the leading operating system vendor
    consider the BIOS time is the local time. Probably the real error was by
    IBM of not putting an RTC in its first PC, then defaulting it to local
    in the XT or AT. But who would have guessed?

    However, MS-DOS wasn't changing the time by itself, whereas nowadays,
    every OS does it automatically, which is a problem with several systems
    installed, and one of them is inheriting from MS-DOS.

    Laurent

    > ----- Forwarded message from Colin Watson -----
    >
    > Subject: Bug#504590: clock-setup: OpenSolaris uses local time by default
    > Reply-To: Colin Watson , 504590@bugs.debian.org
    > Date: Wed, 5 Nov 2008 12:26:56 +0000
    > From: Colin Watson
    > To: submit@bugs.debian.org
    > X-Mailing-List: archive/latest/135870
    > X-CRM114-Status: Good ( pR: 43.0249 )
    >
    > Package: clock-setup
    > Version: 0.97
    > Severity: normal
    > Tags: patch
    > User: ubuntu-devel@lists.ubuntu.com
    > Usertags: origin-ubuntu ubuntu-patch jaunty
    >
    > Apparently OpenSolaris uses local time by default rather than UTC. This
    > was reported as
    > https://bugs.launchpad.net/ubuntu/+s...up/+bug/279065, and,
    > while I found this surprising (as noted in the comments on that bug),
    > the configuration file quoted there contains a comment that seems to be
    > adequate evidence. I suggest the attached patch; comments?
    >



    --
    / Leader de Projet & Communauté | I'm working, but not speaking for
    \ G11N http://fr.opensolaris.org | Bull Services http://www.bull.com
    / FOSUG http://guses.org |




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

  2. Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default

    On Thursday 06 November 2008, Laurent Blume wrote:
    > The executive summary is, yes, by default, on *x86* systems, both
    > Solaris and OpenSolaris considers the system clock to be the local
    > time.


    Thanks a lot for the detailed explanation.

    It means that IMO the current patch proposed by Colin is not correct:
    - we should make sure we only default to local time on x86, and not on
    sparc
    - we probably should ask the user what he actually wants because he
    may be running OpenSolaris in UTC, for example if it's the only
    installed OS and especially if it's co-installed with Linux and not
    Windows

    Looking at os-prober, we seem to detect Solaris (I'm assuming that covers
    OpenSolaris) as Solaris on both arches, but with differing long
    descriptions (Solaris/IA32 and Solaris/SPARC).

    IMO having different long descriptions does not really make sense as the
    user already knows what arch is running (and IA32 would seem to imply
    that OpenSolaris has no 64-bit support?).

    Cheers,
    FJP



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

  3. Bug#504590: (forw) Bug#504590: clock-setup: OpenSolaris uses local time by default

    Frans Pop a écrit :
    > It means that IMO the current patch proposed by Colin is not correct:
    > - we should make sure we only default to local time on x86, and not on
    > sparc
    > - we probably should ask the user what he actually wants because he
    > may be running OpenSolaris in UTC, for example if it's the only
    > installed OS and especially if it's co-installed with Linux and not
    > Windows


    I think that's correct.

    > Looking at os-prober, we seem to detect Solaris (I'm assuming that covers
    > OpenSolaris) as Solaris on both arches, but with differing long
    > descriptions (Solaris/IA32 and Solaris/SPARC).
    >
    > IMO having different long descriptions does not really make sense as the
    > user already knows what arch is running (and IA32 would seem to imply
    > that OpenSolaris has no 64-bit support?).


    I suppose that's only a shortcut (should be x86 instead of IA32, IMO,
    but no big deal :-) ).
    On Solaris SPARC, since version 7, both sparc (32 bit) and sparcv9 (64
    bit) arch are supported.
    On Solaris x86, since version 10, and including OpenSolaris, both i386
    and amd64 arch are supported.
    So from an installer point of view, on a given family of architecture,
    there's no possibility to differentiate between 32/64.

    Laurent
    --
    / Leader de Projet & Communauté | I'm working, but not speaking for
    \ G11N http://fr.opensolaris.org | Bull Services http://www.bull.com
    / FOSUG http://guses.org |



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

+ Reply to Thread