slave crashed about some load

[franky6464]

I crashed the slave after some load 500-600 mbit (170 connections) at about 100 channels. Error message in dmesg "out of memory) the sysctl.conf is adjusted , filelimit increased.
Dedicated Server with 32 GB Ram.

Check the error in attachment file

[MagSat]

Did you restart server after limits.conf edition ?

[franky6464]

i dont have reboot, while i think user@ubuntu:~$ sudo sysctl -p must work.

But when i ´m at home i make reboot and test some load again

[Evileyes]

what kind of streams are you using hls, mpegts..? and with ffmpeg?

[franky6464]

after reboot, a new crash


Debian 9.11

i'm not the transcoding specialist, but my streams are transcoded in h.264 Mpeg4-AVC, so no hls. the output of the lines has m3u8. i had already made a line with these streams. The lines of exchange partners have most ts or mpegts. After increasing the limits a slave just crashed again, this time with another error.

*** Error in `slave': double free or corruption (!prev): 0x00007fe3800b1320 ***
======= Backtrace: =========
/lib/x86_64-linux-gnu/libc.so.6(+0x70bfb)[0x7fe40cd3bbfb]
/lib/x86_64-linux-gnu/libc.so.6(+0x76fc6)[0x7fe40cd41fc6]
/lib/x86_64-linux-gnu/libc.so.6(+0x7780e)[0x7fe40cd4280e]
slave(+0x29bf5)[0x5620b5b9cbf5]
slave(+0x2b02e)[0x5620b5b9e02e]
/lib/x86_64-linux-gnu/libpthread.so.0(+0x74a4)[0x7fe40f86d4a4]
/lib/x86_64-linux-gnu/libc.so.6(clone+0x3f)[0x7fe40cdb3d0f]
======= Memory map: ========
5620b5b73000-5620b5bb5000 r-xp 00000000 09:02 63449769 /usr/local/bin/slave
5620b5db4000-5620b5db5000 r--p 00041000 09:02 63449769 /usr/local/bin/slave
5620b5db5000-5620b5db6000 rw-p 00042000 09:02 63449769 /usr/local/bin/slave
5620b5db6000-5620b5e2e000 rw-p 00000000 00:00 0
5620b6b0a000-5620b6b66000 rw-p 00000000 00:00 0 [heap]
5620b6b66000-5620b6b9d000 rw-p 00000000 00:00 0 [heap]
7fe254000000-7fe257ff9000 rw-p 00000000 00:00 0
7fe257ff9000-7fe258000000 ---p 00000000 00:00 0
7fe258000000-7fe259513000 rw-p 00000000 00:00 0
7fe259513000-7fe25c000000 ---p 00000000 00:00 0
7fe25c000000-7fe25fff9000 rw-p 00000000 00:00 0

[franky6464]

now my second Slave....this one run on ubuntu 18.04,

Server run more than 10 hours.

[58503.301332] Out of memory: Kill process 933 (slave) score 962 or sacrifice child
[58503.301532] Killed process 933 (slave) total-vm:52429524kB, anon-rss:32352044kB, file-rss:0kB, shmem-rss:0kB
[58504.484521] oom_reaper: reaped process 933 (slave), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB

[franky6464]

i have now updated to the new Version from today and will see

[moh6000]

verify your streams one crashe slave

[franky6464]

verify your streams one crashe slave

how can i verify my streams, i come and see the slave are down....i dont see why, i post the dmesg, i dont see more....
there are 1000 channels/Streams on the slave, and its not one slave, the second one crash also, but not fast, the first afer 3 hours running, the second after 8h running...
and so on...

[moh6000]

you verify stream have extention m3u8 and https://