"too large" is mostly a badly named error that indicates that some limit is reached regarding the number of streams that may be used concurrently. Most likely Flash might be leaking streams. Please run "pacmd ls" and attach the output here.

[dwmw2@macbook ~]$ pacmd ls
No PulseAudio daemon running, or not running as session daemon.
[dwmw2@macbook ~]$ ps axf | grep pulse
22747 pts/5    S+     0:00      \_ grep pulse
27004 ?        S<sl   5:11 /usr/bin/pulseaudio --start --log-target=syslog
27062 ?        S      0:00  \_ /usr/libexec/pulse/gconf-helper
[dwmw2@macbook ~]$ play 99.ogg 
ALSA lib pcm_pulse.c:724:(pulse_prepare) PulseAudio: Unable to create stream: Too large

play formats: can't open output file `default': cannot set parameters
[dwmw2@macbook ~]$ killall pulseaudio
[dwmw2@macbook ~]$ play 99.ogg 

99.ogg:

  Encoding: Vorbis        
  Channels: 2 @ 16-bit   Track: 05
Samplerate: 44100Hz      Album: The Best Of Suzanne Vega - Tried And True
Replaygain: off         Artist: Suzanne Vega
  Duration: 00:03:14.00  Title: 99.9F°

In:0.62% 00:00:01.21 [00:03:12.79] Out:53.2k [ -====|====  ] Hd:0.0 Clip:0    ^C
Aborted.
[dwmw2@macbook ~]$ pacmd ls
Welcome to PulseAudio! Use "help" for usage information.
 ... will attach ...

Created attachment 383154 [details]
output from 'pacmd ls'

I had killed and restarted firefox -- shouldn't that have ensured that any flash player streams were also gone?

The interesting "pacmd ls" output is of course when PA is refusing further streams to be created. "pactl list" would be similarly interesting.

Is there any interesting output in syslog?

(In reply to comment #2)
> [dwmw2@macbook ~]$ pacmd ls
> No PulseAudio daemon running, or not running as session daemon.

If pacmd des not find the control socket that is necessary for its work it sends pa a SIGUSR1 which loads the module which provides that socket. It waits 200ms or so for that. if your PA is overloaded because too many streams are created in a loop those 200ms might not suffice, hence try to run pacmd ls a couple of times.

Lots of these... 
Jan 12 02:14:38 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.
Jan 12 02:14:38 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.
Jan 12 02:14:38 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.
Jan 12 02:14:38 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.
Jan 12 02:14:40 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.
Jan 12 02:14:59 macbook pulseaudio[27004]: sink-input.c: Failed to create sink input: too many inputs per sink.


I had run 'pacmd ls' a few times, and got the same result every time. I just showed you one.

I don't know why PA would be overloaded -- this system doesn't really _do_ anything with audio except for the occasional use of flash player (or xine), and when the cats hit the pause/play button mpd will sometimes start playing something.

Maybe some app plays an event sound?

PA seems to think that there are too many streams lingering. The question is, which ones are those. 

Does "pactl list" give any useful output when this problem occurs?

Created attachment 386011 [details]
Output of "pacmd list"

Output of "pacmd list" when my system is having the "Unable to create stream: Too large" error

I just attached an output of the problem in progress on my system. It appears that there are 32 stuck sink inputs for gnome-terminal's error bell. I'm not sure how this happens though because I can't actually seem to make it happen manually, even if I sit in a terminal window and hold down the backspace key to force nonstop error bells.

Created attachment 387749 [details]
Output of pacmd ls

Created attachment 387750 [details]
Output of pactl list

The above outputs of 'pacmd ls' and 'pactl list' are taken when both Totem and MPlayer suffer from this problem.

Hmm, this is weird. I really wonder how I could reproduce this on my system here... The bell sounds seem to e stuck in DRAINED state, which should be impossible...

This is now fixed upstream. Cached samples that were played by libcanberra did get queued up while an audio device was suspended but could not get unsuspended, thus taking up all slots sooner or later.

Will upload a new version soon.

*** Bug 538472 has been marked as a duplicate of this bug. ***

pulseaudio-0.9.21-6.fc13 has been submitted as an update for Fedora 13.
http://admin.fedoraproject.org/updates/pulseaudio-0.9.21-6.fc13

pulseaudio-0.9.21-5.fc12 has been submitted as an update for Fedora 12.
http://admin.fedoraproject.org/updates/pulseaudio-0.9.21-5.fc12

pulseaudio-0.9.21-6.fc13 has been pushed to the Fedora 13 stable repository.  If problems still persist, please make note of it in this bug report.

pulseaudio-0.9.21-5.fc12 has been pushed to the Fedora 12 stable repository.  If problems still persist, please make note of it in this bug report.