[edit: Oops. It looks like SpAwN_gUy wrote a bunch of stuff after I started writing this. I'll get to those responses later]
First off, I'd like to mention that I updated my x264 builds to r634. links:
32-bit 64-bit
Also, I don't know if I mentioned this, but you only need my builds if you are using controller-based encoding or piped input on the agent. If you are using non-piped input to an agent-based encoder, you may use any x264 build you want.
Quote:
Originally Posted by foxius
Yeah, that's exactly the problem I'm having sometimes. One agent is stuck and others are just waiting.
|
Does the agent show any errors? What's the CPU usage like when this happens?
One problem that I have had is that if an agent stops doing anything without informing the controller that something's wrong, then the encoding will never finish. If you restart either the controller or agent it should complete, but there's not really a good answer right now. I have been trying to think of a nice, clean way to handle this since long before I released x264farm...
Quote:
Perhaps it would work even without removing the temp files. I'll try not to delete them next time.
I'm not sure if it's possible to detect a DELFIRST dead loop, but I'd suggest putting a limit to consecutive DELFIRST messages received from an agent. If it's reached, then the controller assumes that the agent is stuch and sends a signal to drop the job and restart. Of course, this is only a quick&dirty workaround. Can't think of anything else.
|
Well... I really don't think DELFIRST is related to the problem you're having. The DELFIRST condition is a product of the controller so doing something with the agents won't get anywhere.
Also, if it gets into a loop, the agents will still be encoding. The loops occur when the controller throws out everything that an agent does and tells the agent to encode the same thing again. As long as --batch is set to larger than x264's --keyint option, I don't think it's possible to get in a loop.
As to what
is causing the problem you're having, I don't really know. Is it doing it a lot, or was it just that one time? If multiple times, was it always the same agent that was having the problem?