-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On Sat, Jun 20, 2015 at 12:27:59PM +0000, Curt wrote: > On 2015-06-20, <to...@tuxteam.de> <to...@tuxteam.de> wrote: > > > > And the script is linked in /etc/rc[06].d to stop. So with SysV, at > > least, things are set up to update the hwclock at shutdown. > > I guess that settles it then. I do find the wiki and debian doc a little > misleading. I do also remember having to us "hwclock" by my lonesome to > set the hardware clock in the past, but maybe that was in another > distribution, or another universe entirely. > > And if you did set the hardclock with 'hwclock' to another time than > the system time, *that time* would not be carried over subsequent > to a reboot, because the script would write the system time to the > hardware clock at shutdown, thus cancelling what you'd done.
Correct. > So apparently I had it exactly backwards, which is almost correct in a > way. Background is that the "hardware clock" is pretty imprecise (the clock of the running system is much better, especially when corrected by NTP). The command hwclock has a provision to correct that (it tries to calculate the hardware clock's drift and to compensate for it at boot (option --adjust). I don't see that the init script is using that, but perhaps it can be set in /etc/default/hwclock. More research needed to know if/how this can be put to advantage. regards - -- tomás -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iEYEARECAAYFAlWFfD8ACgkQBcgs9XrR2kZB+gCfcIy3dWvV9VW60t3jzEh0CHPd lhIAn1EaWwiKS5DnoZW9bWPj77uLImCp =WesO -----END PGP SIGNATURE----- -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org Archive: https://lists.debian.org/20150620144415.ga2...@tuxteam.de
