Re: Questions about GPSD as a time source with Chrony

[Dominick C. Pastore]

Hi Gary,

On 4/5/22 2:13 PM, Gary E. Miller wrote:
> > > > The GPSD Time Service HOWTO seems to suggest using the socket
> > > > interface for Chrony over the shared memory interface.
> > >
> > > Yes.
> > >    
> > > > Is there some
> > > > advantage to the socket interface?
> > >
> > > No.  Just the opposite.
> >
> > If by "just the opposite" you mean there is a disadvantage to the
> > socket interface (as opposed to merely no advantage),
>
> Correct.
>
> > that begs the
> > questions: 1) Why is the SHM interface better,
>
> It is more general (works with ntpd and chrinyd).  It avoids the
> stupid effects of SELinux, AppArmore, etc.  It avoids file permission
> problems.
>
> > and 2) should the
> > HOWTO reflect that reality?
>
> No, too much details turns off readers.

Fair enough. That said, the HOWTO already includes instructions for both the 
socket method and the SHM method. If the SHM method is better, wouldn't it at 
least make sense to put that one first, or otherwise revise the HOWTO so it 
doesn't look like the socket method is preferred?

> > I appreciate the responses.
>
> Be sure to pass on your knowledge to the next newbie.

I would be happy to.

Thanks,
Nick