in frustration wih many segfaults experienced as mentioned in https://github.com/karlheyes/icecast-kh/issues/460 I setup a fresh Debian 12 server and compiled the latest icecast-kh version 2.4.0-kh22, configured it as a slave relay for many mounts, and ran it successfully.
But it didnt last - it too crashed very soon after startup - again more bloody SEGFAULTS SEGFAULTS!
from my kern.log:
2024-09-17T10:52:17.505260+02:00 ingest kernel: [ 3047.448456] show_signal_msg: 1 callbacks suppressed
2024-09-17T10:52:17.505293+02:00 ingest kernel: [ 3047.448461] icecast[12274]: segfault at 20 ip 00007f9ffb700ba0 sp 00007ffe0edefe38 error 4 in libc.so.6[7f9ffb699000+155000] likely on CPU 7 (core 0, socket 14)
2024-09-17T10:52:17.505300+02:00 ingest kernel: [ 3047.448480] Code: 92 0f 84 89 fd ff ff 8d 48 26 83 f8 da 72 93 48 b8 09 00 00 08 2c 00 00 00 48 d3 e8 a8 01 74 82 e9 6b fd ff ff 0f 1f 44 00 00 <8b> 47 10 89 c2 81 e2 7f 01 00 00 a8 7c 75 51 48 83 ec 08 85 d2 75
2024-09-17T13:36:46.833219+02:00 ingest kernel: [12916.033875] icecast[13079]: segfault at 30 ip 00007f65546ff2c0 sp 00007f65516fc928 error 4 in libc.so.6[7f6554699000+155000] likely on CPU 12 (core 0, socket 24)
2024-09-17T13:36:46.833255+02:00 ingest kernel: [12916.033890] Code: 2f 11 00 ba c2 01 00 00 48 8d 35 47 ac 10 00 48 8d 3d 55 ac 10 00 e8 bf 8b fa ff e8 0a be 08 00 66 2e 0f 1f 84 00 00 00 00 00 <8b> 47 10 89 c2 81 e2 7f 01 00 00 83 e0 7c 0f 85 ac 00 00 00 53 48
The same random instability issue crashing icecast-kh is now seen on two servers!
I'm running icecast 2.4.0-kh13 on an older debian server as the same slave relay - exactly the same icecast.xml - and that runs beautifully for months without any flaw - average concurrency at around ~2000 listeners.
Now why would the newest release several versions ahead of that old clunker, be so unreliable flakey?
what to do? drop back several versions in hope of regaining stability?
It's kinda totally backwards - isnt it?
in frustration wih many segfaults experienced as mentioned in https://github.com/karlheyes/icecast-kh/issues/460 I setup a fresh Debian 12 server and compiled the latest icecast-kh version 2.4.0-kh22, configured it as a slave relay for many mounts, and ran it successfully. But it didnt last - it too crashed very soon after startup - again more bloody SEGFAULTS SEGFAULTS!
from my kern.log: 2024-09-17T10:52:17.505260+02:00 ingest kernel: [ 3047.448456] show_signal_msg: 1 callbacks suppressed 2024-09-17T10:52:17.505293+02:00 ingest kernel: [ 3047.448461] icecast[12274]: segfault at 20 ip 00007f9ffb700ba0 sp 00007ffe0edefe38 error 4 in libc.so.6[7f9ffb699000+155000] likely on CPU 7 (core 0, socket 14) 2024-09-17T10:52:17.505300+02:00 ingest kernel: [ 3047.448480] Code: 92 0f 84 89 fd ff ff 8d 48 26 83 f8 da 72 93 48 b8 09 00 00 08 2c 00 00 00 48 d3 e8 a8 01 74 82 e9 6b fd ff ff 0f 1f 44 00 00 <8b> 47 10 89 c2 81 e2 7f 01 00 00 a8 7c 75 51 48 83 ec 08 85 d2 75 2024-09-17T13:36:46.833219+02:00 ingest kernel: [12916.033875] icecast[13079]: segfault at 30 ip 00007f65546ff2c0 sp 00007f65516fc928 error 4 in libc.so.6[7f6554699000+155000] likely on CPU 12 (core 0, socket 24) 2024-09-17T13:36:46.833255+02:00 ingest kernel: [12916.033890] Code: 2f 11 00 ba c2 01 00 00 48 8d 35 47 ac 10 00 48 8d 3d 55 ac 10 00 e8 bf 8b fa ff e8 0a be 08 00 66 2e 0f 1f 84 00 00 00 00 00 <8b> 47 10 89 c2 81 e2 7f 01 00 00 83 e0 7c 0f 85 ac 00 00 00 53 48
The same random instability issue crashing icecast-kh is now seen on two servers!
I'm running icecast 2.4.0-kh13 on an older debian server as the same slave relay - exactly the same icecast.xml - and that runs beautifully for months without any flaw - average concurrency at around ~2000 listeners.
Now why would the newest release several versions ahead of that old clunker, be so unreliable flakey?
what to do? drop back several versions in hope of regaining stability? It's kinda totally backwards - isnt it?