Open litianyu313 opened 1 year ago
add the ff_log_dir ./objs; to your config file
Seems a bug.
There is a way to get out of this loop if you push transcode to another vhost and disable transcode on that vhost example with exec
listen 19356;
max_connections 1000;
http_api {
enabled on;
listen 1985; # Specify the desired port for the HTTP API
crossdomain on;
raw_api {
enabled off;
allow_reload off;
allow_query off;
allow_update off;
}
}
vhost __defaultVhost__ {
exec {
enabled on;
publish /usr/local/bin/ffmpeg -hide_banner -loglevel error -fflags nobuffer -flags low_delay -i [url] -c:v libx264 -b:v 2000k -vf 'scale=1280:720:force_original_aspect_ratio=decrease' -async 1 -vsync 1 -c:a aac -b:a 96k -f flv rtmp://127.0.0.1:[port]/[app]?vhost=transcode.local/[stream]_720p -c:v libx264 -b:v 1200k -vf 'scale=960:540:force_original_aspect_ratio=decrease' -c:a aac -b:a 96k -f flv rtmp://127.0.0.1:[port]/[app]?vhost=transcode.local/[stream]_540p -c:v libx264 -b:v 550k -vf 'scale=640:360:force_original_aspect_ratio=decrease' -c:a aac -b:a 96k -f flv rtmp://127.0.0.1:[port]/[app]?vhost=transcode.local/[stream]_360p;
}
}
vhost transcode.local {
exec {
enabled off;
}
}
Description
Please description your issue here
SRS Version:
5.0.152
SRS Log:
Replay
Please describe how to replay the bug?
Step 1: get srs 5.0.152
Step 2: config file of transcode like this
Step 3: run srs push stream
rtmp://127.0.0.1:1935/live/livestream
by obs, and then more and more ffmpeg process start by ps -ef|grep ffmpeg. And output url become long and long like :rtmp://127.0.0.1:1935/live?vhost=test.com/livestream_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo_logo
Expect
In the early version, it ran okay, but in bug version 5, it seems to have some problems.
TRANS_BY_GPT3