Re: [chrony-users] late-arriving hardware timestamps

[Aaron Ball <aball@xxxxxxxxxxxxxx>]

In general, does delaying processing to get more HW timestamps seem
likely to be a good tradeoff for overall timekeeping quality? I.e., is
there a clear empirical consensus on whether updating ~100 μs later or
missing ~10% of hardware timestamps would have a bigger impact?

On Tue, Nov 23, 2021 at 11:38 AM Miroslav Lichvar <mlichvar@xxxxxxxxxx> wrote:
>
> On Tue, Nov 23, 2021 at 11:21:13AM -0500, Aaron Ball wrote:
> > On our test hosts with X550 NICs, we're seeing fewer hardware
> > timestamps under 4.1 than under 3.2.
>
> There was a workaround added in 3.3 to wait (for up to 200
> microseconds) for late HW TX timestamps of client requests. I'd expect
> 4.1 to be better than 3.2, not worse.
>
> Can you please post your chrony.conf? Nothing else changed, like
> kernel, drivers, firmware?
>
> If you enable debug log (-d -d and chronyd compiled with
> --enable-debug), does the problem disappear? It would be good to
> confirm that you are getting the "Suspended RX processing" messages,
> but you might need to patch the code to log those messages at a higher
> log level and disable the debug output to still be able to reproduce
> the problem.
>
> --
> Miroslav Lichvar
>
>
> --
> To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "unsubscribe" in the subject.
> For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
> with "help" in the subject.
> Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.
>

--
To unsubscribe email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "unsubscribe" in the subject.
For help email chrony-users-request@xxxxxxxxxxxxxxxxxxxx
with "help" in the subject.
Trouble?  Email listmaster@xxxxxxxxxxxxxxxxxxxx.