Urgent issue with network rendering

[nrygpu]

Hello,

I was getting advice on the best method to render a large animation from C4D here: viewtopic.php?f=30&t=60308

I figure it makes sense to post this here as well as I am currently running Octane 3.06 downloaded from this thread:

I have setup C4D on a Windows 10 based system that has Quadro GPU's. The Windows "master" computer is running C4D R18 with latest update, Octane C4D plugin 3.06 and standalone 3.06. Next I have a GPU server with 8 GPU's that is running Ubuntu 16.04 LTS and also Octane standalone 3.06. I have run the daemon script in Linux and it successfully generates the run_octane_daemon.sh

Next I run the run_octane_daemon.sh and everything seems to be working as it should. However, the problem I am having now is that no matter what I try I can't see the render slave on the master. It simply does not show up in either C4D or octane standalone on the master. As you can see in the screenshot below, it is all running correctly except no slave is detected by the master.

https://www.screencast.com/t/YdEooXxqMxcT

I have spent hours checking the network settings. I made sure I can ping to the private IP of both the Windows master and the Linux slave and vice versa. I also opened up firewall settings to make sure that nothing is blocked. There is no virus software running in this testing environment. This is all running inside of an HPC grade data center with 10Gbps networking between the machines.

I am really stuck here because I need to start rendering this animation, but the only way to get it done is if I can use all 20 GPU's via netrender.

Help is very much appreciated!

[bepeg4d]

Hi,
I understand that you are in a super hurry, but opening several discussions with the same question is not the best way, and will not give you the solution faster.

ciao beppe

[nrygpu]

Ah, please excuse me...I was not sure if I was posting in the correct place or not. It is a network rendering issue of course but then I am also using 3.06 TEST version so I was not sure if it might be an issue with that version.

I will follow up with you on the original thread. Thanks

[aoktar]

First try to disable firewall and antivirus to be sure a blockage or not.

[nrygpu]

I made sure that the firewall is open and network settings are not blocking anything. I am fully able to ping each the master from the slave and vice versa.

Please see my responses on the other thread: viewtopic.php?f=30&t=60308&p=308621#p308621

I am really in a bind here. I have contacted OTOY Support as well. I really hope that someone can suggest something that will help otherwise I will have lost this project / customer. Thanks!

[nrygpu]

Hello,

I just found one thing that may be causing an issue. I have setup a totally fresh Ubuntu 16.04 LTS install and when I get to the part of executing the Linux command: sudo ./install-daemon.sh
to install the slave daemon it is showing the following console output:

ubuntu@cosmosgpu16:~/octane$ sudo ./install-daemon.sh
### OctaneRender Slave Daemon installation
please enter a network port for the daemon between 1024-65355
Leave empty for default (48000):
created the daemon startup script /home/ubuntu/octane/run_octane_daemon.sh
./install-daemon.sh: 77: ./install-daemon.sh: bc: not found
./install-daemon.sh: 77: [: !=: unexpected operator
create the desktop file octane_daemon.desktop
copy desktop file for menu inclusion into /home/ubuntu/.local/share/applications
copy desktop file for auto-run into /home/ubuntu/.config/autostart
ubuntu@cosmosgpu16:~/octane$

Does this look correct or am I missing something?

Thanks,

Ry

[bepeg4d]

After several PM the issue seems to be solved, with the exception that the slave is correctly up and running, but is not listed in the master network window.
Probably due to the linux daemon with windows master, I guess
ciao beppe

[nrygpu]

YES!!! It is working...

Here is a command line eye candy shot:

I will try to add the:
--alias "%COMPUTERNAME%"
to the end of my linux command.

Actually I am not using a script on the Linux slave because it did not seem to work properly. I am running this string on the Linux slave:
sudo ./octane_slave --net-master-address 10.0.0.185 --net-master-port 1025

The question now is if I can add another slave with 8 more GPU's while the animation frames are rendering out the still frames, or do I need to stop it all; add the second slave node...then restart the render job?

I will report back on that soon!

Thanks again for all the help. I think the issue is actually octane licensing causing the network render checkbox to freeze up and uncheck itself. I have to be very critical of the procedure I take to: 1. open C4D 2. launch LV to activate octane licensing 3. go into Render pull down > Render Settings > check the Network Render box > click Network Preferences > make sure the box is checked there too. 4. Save the file 5. Only then will network rendering work from PV or Commandline.exe