[pulseaudio-discuss] main.c: daemon startup failed.

Lennart Poettering lennart at poettering.net
Mon Apr 14 09:30:26 PDT 2008

On Sun, 13.04.08 19:53, Sean McNamara (smcnam at gmail.com) wrote:

> > I: main.c: Dude, your kernel stinks! The chef's recommendation today is Linux with high-resolution timers enabled!

> The above line is pretty good advice. No hrtimers means you're likely to
> get dropouts. Very likely. Even if you resolve the problems keeping
> sound from nominally working, it will probably not be enjoyable for you,
> particularly when the system is under moderate CPU or cache load. But,
> the prior problem of getting sound at all still remains... What version
> of ALSA are you running?

This is not quite right. The impact of not having hrtimers is not as
big as this informational message might suggest. It won't cause
drop-outs. At least not on any of the released versions of PA. It's
different in the glitch-free branch, where not having hrtimers will
cause PA to fall back to the the traditional playback model, but even
that won't cause any dropouts. In fact one could even argue that it
would make perfect sense to lift this artificial limitation, since
systems with HZ=1000 are still perfectly capable of providing sched
latencies in the range of a few milliseconds.

As the "I" prefix suggests this messages is purely
informational. Nothing to worry about.

> The source didn't open. The "E" indicates this is an unrecoverable
> error, IIRC... in other words, this is causing PA to exit.
> Can you use the "arecord" program on hw:0? Is there any reason for you
> to suspect that your hardware doesn't support sound input? You haven't
> stated what chipset you're using yet, unless I've just missed it...

E means error. Not unrecoverable error. PA fails if a module fails to
load that is inside a ".fail" section in default.pa.

This might be caused by a simplex audio devices or by one where
certain channel configs cause recording at the same time to fail.

Unfortunately ALSA doesn't hand out very useful error codes. if you
really want to know what's going on it is most of the time better to
use strace and check what the actualy system calls return.

> > D: cli-command.c: Checking for existance of '/usr/lib/pulse-0.9/modules//module-x11-publish.so': success
> > D: module-x11-publish.c: using already loaded auth cookie.
> > X connection to localhost:10.0 broken (explicit kill or server shutdown).
> You should probably comment out module-x11-publish
> from /etc/pulse/default.pa. It's not as useful as it sounds, and
> depending on the circumstances, that X error alone could be causing the
> server to shut down.

In the default config of PA 0.9.10 module-x11-publish doesn't cause
the server satrtup to fail anymore.


Lennart Poettering                        Red Hat, Inc.
lennart [at] poettering [dot] net         ICQ# 11060553
http://0pointer.net/lennart/           GnuPG 0x1A015CC4

More information about the pulseaudio-discuss mailing list