Massive memory footprint from rtmpdump

dinkypumpkin dinkypumpkin at gmail.com
Mon Feb 10 15:19:37 EST 2014


On 10/02/2014 17:39, Nick wrote:
> On Mon, 10 Feb 2014 15:20:12 +0000
> Jimmy Aitken <jimmy.aitken at gmail.com> wrote:
>> Does anyone have any thought as to why this may have started?  I'm now
>> reticent to leave any jobs running automatically via cron in case this
>> starts up again, and my machine in unusable when I get back to it.

This problem has been reported a number of times over the years. 
Apparently, when rtmpdump is fed garbage or suffers dropouts due to 
upstream problems, it can sometimes become a runaway and begin eating 
memory.  Perhaps it's just obeying some bogus parameter values - I don't 
really know.  There was a patch to rtmpdump a couple of years ago 
purporting to address the problem, but it hasn't proved to be a complete 
solution.  I've been hit a few times, but like you I've never been able 
to repeat the error to debug it.

> Isn't that what ulimits are for? Limiting the memory usage by a
> process? Ah, that seems to be for file sizes. I'm sure there is a way to
> limit memory so you can set-up a safety net (at least another script
> monitoring rtmpdump's memory usage and killing it if it gets too big).

Using ulimit to constrain virtual memory use should work on Linux, but 
it never worked for me on OSX.   I don't think that particular ulimit 
setting is obeyed on OSX.  The new virtual memory setting is accepted, 
but seems to have no effect.




More information about the get_iplayer mailing list