Hi Bill, Thanks for the suggestions.
On Thu, Mar 10, 2022 at 6:37 PM Bill Unruh <un...@physics.ubc.ca> wrote: > > Why not just use the prodedure that both Miroslav and I suggested-- run a cron > job once a minute to touch /var/log/chrony.drift. That should be pretty > reasonable even if the internet went down for a couple of hours before the > power outage. Most of the computer clocks do pretty well freewheeling, so that > time should not be too far out (less than a second and certainly not fast by > a minute ) for use when chrony comes back up again and the internet comes up > at powerup so that the count is progressive. I would also make sure that that > program which dies if the time ever goes backwards starts up a few minutes > after chrony comes up again. Yeah this is what I've implemented. At early boot time (pre-chronyd) correct the time to last-known-good if the RTC is behind system time (assuming RTC always falls back). Then, aside from NTP synchronization, on a regular basis update last-known-good time. I must say it works quite well. I did have to fiddle with the "system time provider" in my tool, because during the lifetime of chronyd, /dev/rtc is unavailable for reading. So during the early boot time correction I'm using the RTC from /dev/rtc, then, for runtime last-known-good time updating, I'm using the system-wide date/time as the system time. > Note also since your program presumably writes out files, you could use the > timestamp on those files instead of chrony.drift as the source for you time at > bootup. Yes, I'm using the timestamp of a custom file dedicated for this purpose. Thanks for all the great help! Met vriendelijke groeten / Kind regards, Kris van Rens -- To unsubscribe email chrony-users-requ...@chrony.tuxfamily.org with "unsubscribe" in the subject. For help email chrony-users-requ...@chrony.tuxfamily.org with "help" in the subject. Trouble? Email listmas...@chrony.tuxfamily.org.