[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: No PPS from gpsd

From: Gary E. Miller
Subject: Re: No PPS from gpsd
Date: Thu, 29 Jul 2021 13:40:25 -0700

Yo Jim!

On Thu, 29 Jul 2021 20:16:49 +0000 (UTC)
Jim Pennino <> wrote:

> > Did you use "-D 4" with gpsd?  That is the best way to debug PPS.
> > As gpsd only logs to syslog, I did not do that as setting syslog to
> > debug will fill with messages from everything on the system. IMHO
> > gpsd should have an option to set the logging facility to something
> > quieter such as usr1.  
> Where it logs to is configurable by your logging system.  I only need
> a minutes worth.
> No, where a program logs to is determined by which facility is
> compiled in absent a switch to change it at run time. gpsd is
> obviously compiled to log to LOG_SYSLOG.

Yes, but what your syslogger does with it is up to you.  I don't know
what syslogger you use, so can not provide configuration hint.

> The issue is that putting
> gpsd in debug mode is only part of it. You also have to set the log
> level of the syslog file to a level of LOG_DEBUG which means
> EVERYTHING on the system is writing huge piles of information to the
> syslog file. If you could change the facility to one LOG_LOCAL0
> through LOG_LOCAL7 that would not be an issue as few, if any, other
> things use those facilities.

I only need 60 seconds of debug.  You are the first ever to make this
complaint, and people send gpsd from systemd(umber) to this list almost
every day.

Since it appears to be a sysmtemd(umpkopf) issue, I have no help on it.

> > Why -b?  You are not running on bluetooth.
> > Because the gpsd documentation says it is read-only mode, not that
> > it is read-only mode only for USB and bluetooth.  
> Actually, the doc says:
>         "Broken-device-safety mode, "
> Is your receiver broken?  Otherwise, bad idea.
> What the man page really says is:
> Broken-device-safety mode, otherwise known as read-only mode. A few
> bluetooth and USB receivers lock up or become totally inaccessible
> when probed or reconfigured; see the hardware compatibility list on
> the GPSD project website for details. This switch prevents gpsd from
> writing to a receiver. It does NOT say that This switch prevents gpsd
> from writing to only USB or bluetooth receivers. 

Tweedle-dee, tweedle-dum.  I write it, I know what I meant.  Do'not
use it unless your receiver is broken.  Using it costs you substantial

> > > Now I have no clue where to go next...    
> > 
> > Shoot systemd(umber) in the head.
> > Who is going to pay me to rewrite Ubuntu?  
> Devuan already did.  Many alternatives to Ubuntu that do not use
> systemd(umbest).
> Changing to another OS is not an option and you so far have not shown
> systemd has anything to do with the issue.

Correct, but systemd(umb) usually does end up being the problem.  And
you said it worked when not run under systemd(ribble).  The gpsdebuginfo
output might help.

> The only viable option I
> have right now is to dump gpsd and just use the type 20 driver.

Which costs you a lot of functionality, but does work if all you want is
good time.  Somehow others do not have your problem on Ubuntu woth gpsd
and ntpd.

> + gpsd -V
> gpsd: 3.20 (revision 3.20)

gpsd 3.20 will fail on 23 October 2021.

> /usr/bin/lsof
> + lsof -iTCP:2947 -s TCP:LISTEN
> systemd     1 root  115u  IPv6 1650045 0t0  TCP localhost:gpsd (LISTEN)
> systemd     1 root  117u  IPv4 1648535 0t0  TCP localhost:gpsd (LISTEN)
> gpsd    55519 gpsd    4u  IPv6 1650045 0t0  TCP localhost:gpsd (LISTEN)
> gpsd    55519 gpsd    5u  IPv4 1648535 0t0  TCP localhost:gpsd (LISTEN)

This bug has been seen before.  Somehow both gpsd AND systemd(umbass) are
both on port 2947.  No idea what it means, but not common.

> + command -v aa-status
> /usr/sbin/aa-status
> + aa-status
> apparmor module is loaded.

You said apparmor was off.  Ubuntu apparmor is source of many known problems.

> 5 profiles are in complain mode.
>    /usr/sbin/gpsd

Complain only mode.  Has it complained?  That might shed light on some
other bugs we have been hearing about.

> 6 processes are in enforce mode.
>    /usr/sbin/ntpd (55559)

Not good, but unrelated.

Which leads back to needing the first minute of start up logs.  "-D 4"
is a good start.

Gary E. Miller Rellim 109 NW Wilmington Ave., Suite E, Bend, OR 97703  Tel:+1 541 382 8588

            Veritas liberabit vos. -- Quid est veritas?
    "If you can't measure it, you can't improve it." - Lord Kelvin

Attachment: pgpQRCpTGAIS7.pgp
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]