Icecast dev - Aug 2004 - Bug in ices, playlist mode (ices kh47, libshout kh22)

Hello, I will take your last version, but I have fresher news about the bug that
occurs every mornings in fact...
I think this is ntpdate which is causing the problem, as the time of the machine
change for more than one second.
Here is all the info:
 09:46:17 up 1 day,  1:22,  2 users,  load average: 0.60, 0.71, 0.68
86 processes: 84 sleeping, 2 running, 0 zombie, 0 stopped
CPU states:  10.4% user,   2.3% system,   0.0% nice,  87.3% idle
Mem:    264064K total,   258680K used,     5384K free,   137484K buffers
Swap:   128484K total,        0K used,   128484K free,    60028K cached

  PID USER     PRI  NI  SIZE  RSS SHARE STAT %CPU %MEM   TIME COMMAND
11461 root      18   0  2080 2080  1128 S    15.7  0.7  31:16 ices
11365 root      15   0   972  972   748 R     8.7  0.3   0:00 top
11463 nobody    10   0  2568 2568  1084 S     4.3  0.9   4:49 icecast
11368 root      17   0   508  508   428 S     0.8  0.1   0:00 sleep
    1 root       0   0   484  484   424 S     0.0  0.1   0:07 init

<p>root     11458  0.0  0.7  8168 2080 ?        S    00:00   0:00  \_ ices
/usr/radio/tmp/ices_conf.8007.xml
root     11460  0.0  0.7  8168 2080 ?        S    00:00   0:00  |   \_ ices
/usr/radio/tmp/ices_conf.8007.xml
root     11461  5.3  0.7  8168 2080 ?        S    00:00  31:30  |       \_ ices
/usr/radio/tmp/ices_conf.8007.xml
root     11462  0.0  0.7  8168 2080 ?        S    00:00   0:24  |       \_ ices
/usr/radio/tmp/ices_conf.8007.xml

<p>/etc/cron.daily/ntpdate:
25 Nov 06:23:34 ntpdate[1067]: step time server 212.27.32.176 offset -1.390927
sec

and in fact, ices is working to produce mass logging as
[2003-11-25  10:16:19] EROR playlist-builtin/write_ogg_data failed buffer
allocation
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304924648), resetting timer
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304949773), resetting timer
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304949774), resetting timer
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304949774), resetting timer
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304949774), resetting timer
[2003-11-25  10:16:20] WARN input/input_sleep Sleeping for over 1 second
(304949774), resetting timer

and rotating ices.log!

It is with passthru enable. But ices is not reencoding, just writing and writing
logging... 8Mo per 15 minutes...
the problems started exactly at the time of cron, and the only think suspect in
cron is ntpdate (we have cpu graph with mrtg)

Chris
--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to
'icecast-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

Hello,
ices kh48 seems to have some trouble...
See the stranges characters avec metadata update. (before, it was written bad
address)
and also, it is not connecting to icecast: socket error, ...
I go back to kh47 for the moment.
Chris

[2003-11-25  10:45:50] INFO ices-core/main Streamer version IceS 2.0-kh-pre48
[2003-11-25  10:45:50] INFO ices-core/main libshout version 2.0-kh22 
[2003-11-25  10:45:50] INFO ices-core/main realtime scheduling has been enabled
[2003-11-25  10:45:50] DBUG stream/start_runners starting runner
[2003-11-25  10:45:50] INFO stream/ices_runner Runner 1 ready
[2003-11-25  10:45:51] INFO input-pcm/pcm_initialise_module Initialising PCM
input module
[2003-11-25  10:45:51] DBUG input-pcm/pcm_initialise_module Module PCM using
buffers of 4096 samples
[2003-11-25  10:45:51] DBUG input/initialise_input_modules Module 1 (pcm) has 20
buffers
[2003-11-25  10:45:51] DBUG input/open_next_input_module checking module 1
[2003-11-25  10:45:51] INFO input-pcm/pcm_open_module Retrieving PCM to skip
[2003-11-25  10:45:51] INFO input-pcm/pcm_open_module Retrieving PCM from
standard input
[2003-11-25  10:45:51] DBUG input-pcm/pcm_read metadata updates flagged
[2003-11-25  10:45:51] WARN metadata/metadata_thread_signal Failed to open file
(null) for metadata update:
~<D0>^O@<88>r^<E8><B6>^<F8><95>^G<FE>^
[2003-11-25  10:45:52] INFO stream/process_encode_init Restarting encoder for
PCM input on stream 1
[2003-11-25  10:45:52] INFO audio/downmix_initialise Enabling stereo->mono
downmixing
[2003-11-25  10:45:52] INFO audio/resample_initialise Initialised resampler for
1 channels, from 44100 Hz to 19404 Hz
[2003-11-25  10:45:52] INFO encode/encode_setup Encoder initialising with
bitrate management: 1 channels, 19404 Hz, minimum bitrate 0, nominal 25000,
maximum 0
[2003-11-25  10:45:52] DBUG stream/_output_oggpacket seen new stream, better get
headers
[2003-11-25  10:45:52] DBUG stream/_output_oggpacket samplerate is 19404,
channels is 1
[2003-11-25  10:45:55] DBUG om_shout/check_shout_connected Time we started
stream on SERVER_IP/radio-bro-gwened.ogg
[2003-11-25  10:45:55] INFO om_shout/check_shout_connected Connected to server:
SERVER_IP/radio-bro-gwened.ogg
[2003-11-25  10:45:55] DBUG om_shout/output_ogg_shout initialising output stream
[2003-11-25  14:59:54] EROR om_shout/output_ogg_shout Failed to write to
SERVER_IP/radio-bro-gwened.ogg (Socket error)
[2003-11-25  14:59:54] DBUG om_shout/_output_connection_close closed shout
connection
[2003-11-25  14:59:57] DBUG om_shout/check_shout_connected Time we started
stream on SERVER_IP/radio-bro-gwened.ogg
[2003-11-25  14:59:57] EROR om_shout/check_shout_connected Failed to connect to
SERVER_IP/radio-bro-gwened.ogg (Login failed)
[2003-11-25  14:59:57] DBUG om_shout/_output_connection_close closed shout
connection
[2003-11-25  15:00:00] DBUG om_shout/check_shout_connected Time we started
stream on SERVER_IP/radio-bro-gwened.ogg
[2003-11-25  15:00:00] INFO om_shout/check_shout_connected Connected to server:
SERVER_IP/radio-bro-gwened.ogg
[2003-11-25  15:00:00] DBUG om_shout/output_ogg_shout initialising output stream
[2003-11-25  15:00:25] WARN input-pcm/wait_for_pcm Timeout reading from input
[2003-11-25  15:00:25] INFO input-pcm/pcm_read No more data available
[2003-11-25  15:00:25] INFO input/input_loop Closing input module
[2003-11-25  15:00:25] INFO input-pcm/pcm_close_module Closing PCM input module
[2003-11-25  15:00:25] DBUG input/open_next_input_module checking module 1
[2003-11-25  15:00:25] INFO input-pcm/pcm_open_module Retrieving PCM to skip
[2003-11-25  15:00:25] INFO input-pcm/pcm_open_module Retrieving PCM from
standard input
[2003-11-25  15:00:25] DBUG input-pcm/pcm_read metadata updates flagged
[2003-11-25  15:00:25] WARN metadata/metadata_thread_signal Failed to open file
(null) for metadata update:
~<D0>^O@<88>r^<E8><B6>^<F8><95>^G<FE>^
[2003-11-25  15:00:30] WARN input-pcm/wait_for_pcm Timeout reading from input
[2003-11-25  15:00:30] INFO input-pcm/pcm_read No more data available
[2003-11-25  15:00:30] INFO input/input_loop Closing input module
[2003-11-25  15:00:30] INFO input-pcm/pcm_close_module Closing PCM input module
--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to
'icecast-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.

On Tue, 2003-11-25 at 09:30, iceuse@kezako.net wrote:
> Hello, I will take your last version, but I have fresher news about the bug
that occurs every mornings in fact...
> I think this is ntpdate which is causing the problem, as the time of the
machine change for more than one second.
> It is with passthru enable. But ices is not reencoding, just writing and
writing logging... 8Mo per 15 minutes...
> the problems started exactly at the time of cron, and the only think
suspect in cron is ntpdate (we have cpu graph with mrtg)
ah, yes, big time jumps do affect timing critical apps in a big
way. Check with the kh48 I have on my dir as that has the timer
reset code fixed.

The time jump should be logged but only once, the timer resets to
the new time and continues.   If you don't want the big jump in
the clock then you're best having the ntpdate called more frequently.

karl.

<p>--- >8 ----
List archives:  http://www.xiph.org/archives/
icecast project homepage: http://www.icecast.org/
To unsubscribe from this list, send a message to
'icecast-dev-request@xiph.org'
containing only the word 'unsubscribe' in the body.  No subject is
needed.
Unsubscribe messages sent to the list will be ignored/filtered.