We run icecast-2.4.0-kh22 on kubernetes and are relaying 3 Radio Stations, while 2 of them are running smooth without any issues for weeks, we have troubles with one station.
in the logs with loglevel DEBUG we can's see why its crashing we just get the core dump message, see below.
somewhere i was reading that it's possible to compile with debug flags, is this the way to go?
if we add debug flags, should we keep loglevel also on debug or can we go back to info to have "less clutter" in the log?
would it be of value to store the core dump somehow, we don't have experience on how todo this on kubernetes yet :-)
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG source/source_update_settings on_demand set
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG stats/process_source_stat new node on /radiobeo_iphone.mp3_fb "authenticator" (url)
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG stats/process_source_stat new node on /radiobeo_iphone.mp3_fb "genre" (various)
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG stats/process_source_stat new node on /radiobeo_iphone.mp3_fb "server_name" (Unspecified name)
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG stats/process_source_stat new node on /radiobeo_iphone.mp3_fb "public" (0)
2024-08-28 21:02:54: [2024-08-28 19:02:53] DBUG stats/process_source_stat new node on /radiobeo_iphone.mp3_fb "listener_peak" (0)
2024-08-28 21:02:54: [2024-08-28 19:02:53] INFO source/source_apply_mount Applying mount information for "/radiobeo_iphone.mp3_fb"
2024-08-28 21:02:55: /start.sh: line 34: 18 Segmentation fault (core dumped) icecast -c /etc/icecast.xml
Hi
We run icecast-2.4.0-kh22 on kubernetes and are relaying 3 Radio Stations, while 2 of them are running smooth without any issues for weeks, we have troubles with one station.
The Pod's are Crashing with
in the logs with loglevel DEBUG we can's see why its crashing we just get the core dump message, see below. somewhere i was reading that it's possible to compile with debug flags, is this the way to go? if we add debug flags, should we keep loglevel also on debug or can we go back to info to have "less clutter" in the log? would it be of value to store the core dump somehow, we don't have experience on how todo this on kubernetes yet :-)