[FFmpeg-user] http input issues against S3

[Brian Quandt]

When using http as an input "-i" it seems that error handling (or retrying)
seems weak.

Are there any options to the command line to improve error handling /
retry's on http connections (when used as input).

For instance, if I happen to try to use ffmpeg to transcode from an S3
source, which happens to be on a very heavily used connection, I often get
HTTP 500 type errors (ie 502, 503, etc).   These are nearly never fatal,
but simply indicate a retry is necessary.     ffmpeg using default
parameters, results in messages like:

[https @ 0x1bb2a00] HTTP error 502 Bad Gateway2kbits/s
[mov,mp4,m4a,3gp,3g2,mj2 @ 0x1bb2220] stream 1, offset 0x2072adc0: partial
file
https://api-content.dropbox.com/1/files/auto//of_god_and_dogs_1280x720.mp4:
Invalid data found when processing input

And data is then lost (ie seems like the connection is terminated).

Is there a patch to this behavior or command lines options to set to
improve retry's?

Here's my example ffmpeg command line I'm currently using:

ffmpeg -y -user-agent autodcp/1.0 -headers Authorization: Bearer
nta1234123444123Pr3tp82_S-9wQTdI1?  -i https://somewhere/o.mp4 -f wav
/tmp/out.wav