Sorry for posting a support thread but I am going crazy here....
I have a comp that, only during the render, creates this mess:
I can't reproduce it within the project itself, ram previewing and whatnot. This only occurs during the render.
I have tried:
- Restarting AE
- Double checking that RSMB isn't used (its for sure not)
- Triple checking that RSMB isn't used by using the find effects script
- Completely uninstall RSMB
- Restarting the computer
- Rendering on a different machine
- Making sure all image sequences are the same (interpreted as 23.796)
- Making sure all preComps are of the same frame rate
- Deleting all contents on the hard drive dedicated to AE cache
- Formatting the hard drive dedicated to AE cache
- Turning off all frame blending / interpolation
- Rendering as a different codec
- Rendering within a different wrapper (AVI instead of QT)
- Rendering to a new drive
All with no result. I'm going crazy here! Again.. This does not happen within the timeline as I scrub through it. For whatever reason it only happens during the render, on the same frames. To me, it looks like a common RSMB being dumb problem but there are no instances of RSMB in there.
I am in between gigs right now and have decided to invest my time this week into figuring out with our network technician how to get EC2 to work well with R15. Sadly, the ability is looking fairly bleak to us right now.
Firing up an EC2 instance with Team Render Client installed, then typing the IP in manually on from my workstation seems to work fine, but that requires a TON of configuration for each machine every time I want to use it. A new security group could be created and applied throughout but I am just looking for something cleaner.
On an R14 AMI, I have it set to our office static IP, which has the proper port forwarding throughout our network to reach the net render server. So... You just launch _X_ number of instances in the Amazon control panel and wait for _X_ number of machines to appear within the net render client list. Easy.
But now... since traffic is essentially backwards, I need to somehow get my workstations to dynamically load the IPs of the EC2 instances. Bonjour... is really complicating things. Apparently, you can't really punch holes in a firewall to forward traffic across bonjour ports. Its weird, and the traffic happens at the upper gateway level. I have been testing all day, scouring the internet for solutions and all I can find is a scarily hack-ish method within OS X involving sharing your machine's local broadcast via a dynamic DNS. Amazon doesn't offer OS X instances, so I tried wiring up a Windows equivalent with no luck.
So.... I'm just taking a breather for a second and asking if anybody has had any luck, and if anybody has other creative solutions to this problem. I was thinking, maybe, some type of applet could be made for windows that when launched at startup, phones to a central server detailing it's IP address. Then, some type of script within C4D could batch add those IPs into the team render pool. Still, a pain... But much easier than manually configuring and logging into each instance to add it to the pool.
Any ideas or shared experiences would be appreciated. Even if you have experience in getting airplay-capable devices to share via bonjour across the internet with non-apple devices would be helpful, because in a nutshell, that is exactly the type of network connectivity we are trying to configure.