Re: ntpd vs ntpdate with no hardware clock

On Jul 7, 2024, at 11:03, Warner Losh <[email protected]> wrote:

> On Sun, Jul 7, 2024, 11:54 AM Mark Millard <[email protected]> wrote:
>> On Jul 7, 2024, at 10:23, bob prohaska <[email protected]> wrote:
>> 
>> > On Sun, Jul 07, 2024 at 11:18:56AM -0600, Warner Losh wrote:
>> >> On Sun, Jul 7, 2024, 11:16 AM Ronald Klop <[email protected]> wrote:
>> >> 
>> >>> I created fakertc for my rpi4.
>> >>> 
>> >>> https://www.freshports.org/sysutils/fakertc/
>> >>> 
>> >>> Saves the time on shutdown and sets it back early at boot.
>> >>> 
>> >>> Plus I use ntpdate together with ntpd. Works fine.
>> >>> 
>> >> 
>> >> Curious why the root mod time isn't firing... it whould alrwady do that
>> >> 
>> > Root mod time seems fairly laggy:
>> > rprohask@www:~ % date
>> > Sun Jul  7 10:20:38 PDT 2024
>> > rprohask@www:~ % ls -al /
>> > total 97
>> > drwxr-xr-x  20 root wheel     512 Jul  5 13:21 .
>> > drwxr-xr-x  20 root wheel     512 Jul  5 13:21 ..
>> > unless I'm misusing ls -al, of course....
>> 
>> I mount with noatime in use. Do you?
> 
> Doesn't matter since it's in the superblock.

Interesting. I'd not thought of such.

>> Are you trying to show:
>> 
>> time when file was created (-Ul)
>> time when file status was last changed (-cl)
>> time when file was last modified (-l)
>> time of last access (-ul)
>> 
>> You implicitly specified "last modified". So
>> when was the last change to the root directory
>> representation? It likely is not modified often.
>> (Modifications to file content and subdirectories
>> would not modify / of itself.)
> 
> It's in the superblock. Time of last unmount / update (I'd recalled incorrectly).
> 
>> linux can have relatime vs. strictatime modes.
>> relatime mode updates the atime less often, via
>> a rule set. A linux can have relatime by default.
> 
> Yea... none of that matters...
> 

Well, even with ports/packages, FreeBSD does not have
enough tools to well maintain a RPi4B. For example: for
doing EEPROM updates.

This can lead to also having linux boot media that could
separately suffer the same sort of issues. I have boot
media for RaspiOS64lite (my abbreviation) that I update
in order to in turn do things like update the EEPROM
content of the RPi4B's, for example.

Thus, for the type of overall context (RPi4B), I
considered the notes potentially relevant and, so,
appropriate.

===
Mark Millard
marklmi at yahoo.com