[HACKé]"setting real time clock as..." qui ne se

j’ai déplacé un disque d’une machine à une autre, il boote bien, mais s’arrète sur un message de MàJ de la RTC.
Quelqu’un sait comment contourner ça ?
je peux booter sur init=/bin/sh, si je sais quoi desactiver…

c’est quoi, c’est une synchronisation de l’horloge de temps via internet?

oups désolé :blush:

tiens j’ai trouvé de la doc si ca peut t’aider

[quote]Tunable parameters
The behavior of the high-precision system clock is controlled by a set of tunable parameters located in the /etc/conf/pack.d/clock/space.c file.

update_rtc
If set to 1 (default), the realtime clock (RTC) is set to the new system time after an adjtime(S) adjustment. If set to 0, adjtime( ) does not affect the RTC; SCO does not recommend disabling update_rtc except in certain specific testing situations.

update_rtc_interval
Maximum interval (in seconds) at which the RTC is set to the new system time if update_rtc is not disabled; default is 20.

track_rtc
If set to 1 (default), the system clock tracks the realtime clock to maintain accuracy. If set to 0, the system clock does not track the realtime clock. This variable is automatically set to 0 by the RTC driver if the RTC’s VRT bit indicates that RTC/CMOS might be invalid. This can happen, for example, if the CMOS battery fails.

The kernel ignores the track_rtc variable when ntpd(ADMN) (or any process that calls adjtime(S) every second or so) is running, as if track_rtc were set to 0.

For SCO OpenServer Release 5.0.5 and earlier releases, you must manually set track_rtc to 0 when running NTP. For SCO OpenServer Release 5.0.6 and later, do not adjust the value of track_rtc for NTP. The only reason to set track_rtc to 0 is if the RTC is known to be disfunctional.

track_rtc_interval
Interval (in seconds) at which to update the RTC; default is 20.

clock_drift
The maximum rate (expressed as nanoseconds per second) at which adjtime(S) adjustments are made to the system clock. The default value is 250,000,000 (one quarter-second per second).

tsc_clock_hz
Set to 0 (default) to automatically calculate the HZ rate of the Pentium CPU’s TSC register. SCO does not recommend modifying this parameter.

The following tunable parameters can be used to disable certain features of the high-precision system clock on SCO OpenServer Release 5.0.6 and later releases:

disable_tsc_clock
Set to 1 to use the PIT-based system clock rather than the TSC. If this parameter is set to 1, disable_hires_clock and disable_short_timers are ignored; their functionality is disabled automatically.

disable_rtc_check
Set to 1 to check the RTC’s battery failed'' indicator bit. In SCO OpenServer Release 5.0.6 and later releases, the kernel checks that bit each time it tries to track the RTC. If thebattery failed’’ bit is set, it means that the RTC has determined that the CMOS backup battery is failing. When the kernel sees that, it logs the condition in the /usr/adm/messages file prints a warning to the console, and sets track_rtc to 0 to prevent the system software clock from trying to track a bogus RTC clock. SCO does not recommend modifying this parameter.

disable_hires_clock
Set to 1 to simulate the old 10ms resolution. This affects the time values returned by gettimeofday( ) and similar functions.

disable_pit_read
Set to 1 to handle some PITs that are known to be defective.

disable_short_timers
Set to 1 to disable the short timers. Short timers offer less latency, but cause more hardware interrupts. This parameter affects delays for calls such as select( ) and nap( ).
[/quote]

Oui, je pense que c’est ça.
En fait, j’ai desactivé, en bootant avec un argument init=/bin/bash passé au kernel, le lancement des scripts hwclock.sh et hwclockfirst.sh
Maintenant, ça boote et ça me suffit. J’afinerais plus tard.
MERCI.

apparament c’esr un peut risqué cette histoire, enfin les gens sur le net n’en disent pas des choses rassurantes.

Mais bon qui ne tente rien n’a pas tout!

tes deux machines ont le m^ bios ?