[Icecast] ices2 routinely abandons one of its streams

Karl Heyes karl
Sun Jul 25 15:14:49 PDT 2004

<1089805820.22939.21.camel at bogus.hackers.club>
<20040715084719.GA2698 at hyrule.dyndns.org>
<20040724172233.GA19632 at hyrule.dyndns.org>
Message-ID: <1090793688.14906.27.camel at bogus.hackers.club>

On Sat, 2004-07-24 at 18:22, Frederic Briere wrote:

> So far so good, I haven't seen this problem again in over a week.
> However, it seems that ices2 is ignoring my reconnectattempts=-1.
> Here's an excerpt from my logs:
>   [2004-07-20  18:15:33] EROR om_shout/output_ogg_shout Failed to write to ice.imars.net:8000/cigr.ogg (Socket error)
>   [2004-07-20  18:15:33] DBUG om_shout/_output_connection_close closed shout connection
>   [2004-07-20  18:15:43] DBUG om_shout/check_shout_connected Time we started stream on ice.imars.net:8000/cigr.ogg
>   [2004-07-20  18:15:49] WARN input-oss/oss_read will skip input for a short time
>   [2004-07-20  18:15:58] INFO input/input_loop Closing input module
>   [2004-07-20  18:15:58] INFO input-oss/oss_close_module Closing OSS module
>   [2004-07-20  18:15:58] DBUG input/open_next_input_module checking module 1
>   [2004-07-20  18:15:58] INFO input-oss/oss_open_module Opened audio device /dev/dsp at 2 channel(s), 44101 Hz
>   [2004-07-20  18:15:58] WARN input-oss/oss_read will skip input for a short time
>   (This repeats a couple of times)

> The "too many failures" is the part that bugs me; my requirements are
> that if the connection goes down for any amount of time, then the client
> must retry indefinitely so that it connects as soon as the connection
> comes up again.  Otherwise, I get an urgent page during my vacation. :/

usually when this sort of thing happens, the machine is under some load
eg updatedb.  Try enabling realtime (requires starting as root) but it
drops to the stated user afterwards

The failed to write case is purely some delay (typically network)
sending to icecast.  The input skip bit is ices preventing itself from
eating memory indefinitely.

Send me the ices.log so I can check what led up to the 'Too many
failures' message.


More information about the Icecast mailing list