it was a nightmare I have no space of my hdd when I was looking for the reason du -h messages 760M messages
it's flooded with
Jan 18 15:36:50 localhost kernel: imklog 3.21.9, log source = /proc/kmsg started. Jan 18 15:36:50 localhost rsyslogd: [origin software="rsyslogd" swVersion="3.21.9" x-pid="1969" x-info="http://www.rsyslog.co m"] restart Jan 18 15:43:23 localhost pulseaudio[3014]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers. Jan 18 15:43:23 localhost pulseaudio[3014]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers. ... Jan 20 04:00:43 localhost pulseaudio[3021]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers.
what is the reason, an update in alsa or pulse or most likely because I edited /etc/pulse/default.pa by to add
load-module module-hal-detect tsched=0
https://fedoraproject.org/wiki/Features/GlitchFreeAudio
to get rid of the skips in the sound
I have removed tsched=0 from /etc/pulse/default.pa and rebooted and I still got those entries in messages
Jan 20 04:28:39 localhost pulseaudio[3035]: module-alsa-sink.c: Increasing wakeu p watermark to 40.00 ms Jan 20 04:29:50 localhost pulseaudio[3035]: module-alsa-sink.c: Increasing wakeu p watermark to 80.00 ms Jan 20 04:31:02 localhost pulseaudio[3035]: module-alsa-sink.c: Increasing wakeu p watermark to 160.00 ms Jan 20 04:31:26 localhost pulseaudio[3035]: module-alsa-sink.c: Increasing wakeu p watermark to 176.60 ms Jan 20 04:32:10 localhost pulseaudio[3035]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there was actually nothing to write! Most l ikely this is an ALSA driver bug. Please report this issue to the PulseAudio dev elopers. Jan 20 04:32:10 localhost pulseaudio[3035]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there was actually nothing to write! Most l ikely this is an ALSA driver bug. Please report this issue to the PulseAudio dev elopers.
and worse the sounds skips again
On Tue, 20.01.09 04:17, Muayyad AlSadi (alsadi@gmail.com) wrote:
Jan 20 04:00:43 localhost pulseaudio[3021]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers.
what is the reason,
Your ALSA driver is broken.
That said, I didn't expect that this message would be printed that often on the setups where they happen. PA probably should have some kind of rate limiter on that. However I think it is mostly a DoS in rsyslog that no kind of rate limiting is done there:
https://bugzilla.redhat.com/show_bug.cgi?id=478912
an update in alsa or pulse or most likely because I edited /etc/pulse/default.pa by to add
load-module module-hal-detect tsched=0
Lennart
On Tue, Jan 20, 2009 at 04:10:54AM +0100, Lennart Poettering wrote:
On Tue, 20.01.09 04:17, Muayyad AlSadi (alsadi@gmail.com) wrote:
Jan 20 04:00:43 localhost pulseaudio[3021]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers.
what is the reason,
Your ALSA driver is broken.
That said, I didn't expect that this message would be printed that often on the setups where they happen.
What do you mean by "often"? # grep -c "Most likely this is an ALSA driver bug" messages* messages:1 messages-20081228:16 messages-20090104:5 messages-20090111:7 messages-20090118:13
On my Thinkpad T400 with Centrino 2 and echo 5 > /sys/module/snd_hda_intel/parameters/power_save echo 1 > /sys/module/snd_hda_intel/parameters/power_save_controller
On Sun, 25.01.09 23:12, Tomasz Torcz (tomek@pipebreaker.pl) wrote:
On Tue, Jan 20, 2009 at 04:10:54AM +0100, Lennart Poettering wrote:
On Tue, 20.01.09 04:17, Muayyad AlSadi (alsadi@gmail.com) wrote:
Jan 20 04:00:43 localhost pulseaudio[3021]: module-alsa-sink.c: ALSA woke us up to write new data to the device, but there wa s actually nothing to write! Most likely this is an ALSA driver bug. Please report this issue to the PulseAudio developers.
what is the reason,
Your ALSA driver is broken.
That said, I didn't expect that this message would be printed that often on the setups where they happen.
What do you mean by "often"? # grep -c "Most likely this is an ALSA driver bug" messages* messages:1 messages-20081228:16 messages-20090104:5 messages-20090111:7 messages-20090118:13
That's more how often I expected this message to be printed. However, some folks complain about gigabytes of data like this. And that's what I didn't expect.
But still, I think the biggest issue here is that rsyslog doesn't enforce any kind of rate limitting and is happy to let random users fill up /var.
This is a first-class DoS. Simply do a "cat /dev/urandom | strings | logger" and you can make the whole system go bonkers. And it won't even be accounted to your own disk quota. Yay.
Lennart
-
Lennart Poettering -
Muayyad AlSadi -
Tomasz Torcz