Unreal Horde, Build Accelerator, AWS EC2

Development Platform & Builds
,
1

Hello!

I’m using Unreal Horde Server on a EC2 Instance, and a Horde Agent on a EC2 instance. As far as I understand, this is similar to the official Epic method of using Horde.

When I enable Build Accelerator locally, my local workstation correctly contacts the Horde Server for an Agent to assist. However, the Horde Server provides the private EC2 IP for the Agent. The messaging between my local workstation and the Agent of course fails, because the provided IP is a private IP, not the public one.

How can I reasonably fix this? I may be able to edit Horde source code, or edit some configuration files in unexpected ways, but I feel like I’m missing an easier solution. My setup is similar to the official one, so I suspect there’s a better way to resolve this issue. There doesn’t seem to be a way to edit agent configurations through Horde, they seem to be view only.

Thank you.

2

Did you find any simple solution for this? Agent ‘ComputeIp’ indeed is the private IP and workstation tries to connect to that.

3

There is no simple solution. You can just go into the configs and modify the values manually. This isn’t great for scale, but it should work.

It seems like, in the future (UE 5.5-5.6), Horde will support “relay nodes” which also solve this problem.