Hey!
I am trying to get Unity Accelerator running inside a Linux Container / LXC on a Proxmox host. The container is priviledged, Fuse is enabled in Proxmox (it would not even install otherwise).
OS is Ubuntu Ubuntu 22.04.1 LTS, installation was ran as root, the service is configured to run as user unity-accelerator.
I am getting these errors in the logs:
Wrun exited with error: wrun execute returned errors: [no known AgentID; perhaps run with an invite token?]
{"level":"error","ts":"2022-09-22T09:48:47Z","msg":"Unity Accelerator run error","error":"exit status 2","args":["/opt/unity-accelerator/unity-accelerator","tool","wrun","--persist","/opt/unity-accelerator/storage"],"filename":"/opt/unity-accelerator/unity-accelerator","process":"Unity Accelerator service monitor"}```
The Unity Accelerator shows up as running:
```systemctl status unity-accelerator
* unity-accelerator.service - Unity Accelerator Service
Loaded: loaded (/etc/systemd/system/unity-accelerator.service; enabled; vendor preset: enabled)
Active: active (running) since Thu 2022-09-22 09:48:46 UTC; 3h 7min ago
Main PID: 3688 (unity-accelerat)
Tasks: 10 (limit: 19061)
Memory: 58.5M
CPU: 22.700s
CGroup: /system.slice/unity-accelerator.service
`-3688 /opt/unity-accelerator/unity-accelerator run --persist /opt/unity-accelerator/storage
Sep 22 09:48:46 unity-accelerator systemd[1]: Started Unity Accelerator Service.```
Unfortunately, no ports are open/being listened to:
```netstat -tulpn | grep LISTEN
tcp 0 0 127.0.0.53:53 0.0.0.0:* LISTEN 3459/systemd-resolv
tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN 413/master
tcp6 0 0 ::1:25 :::* LISTEN 413/master
tcp6 0 0 :::22 :::* LISTEN 1/systemd```
Any help would be appreciated.
Howdy QAMightandDelight,
So the error “no known AgentID; perhaps run with an invite token” is due to a missing config parameter. I was able to repro your observations by editing the config file and deleting the value for the AgentID field (so it looked like this: “AgentID”: “oldman-desktop2_id”), then restarting the service.
I’m not sure how or why your config file seems to be missing this parameter. To find your config file, look in the data directory. Based on the logs you shared, I would expect your config file to be here:
/opt/unity-accelerator/storage/unity-accelerator.cfg
The default value for AgentID is your <hostname + ‘_id’>, so in the case of the machine I’m on currently, the hostname is oldman-desktop and my AgentID is oldman-desktop_id.
I attached a copy of a config file with just default values in it, fresh from an install. A couple of UUIDs are x’d out, but otherwise you should be able to use this file to compare with your current config file.
First, I’d make sure you have the AgentID field set, but then I’d make sure that other fields are set with reasonable values also.
Then, if a lot of things are different between this default config and what you have, then I suggest reinstalling and choosing the most basic install instructions:
- Existing Configuration Found, choose Replace…
- Protocol Configuration, choose Accelerate Asset Import Pipeline; uncheck Accelerate Unity Collaborate; leave Registration Key blank
- Security Configuration: Allow insecure access with NO encryption
Make sure to restart the service after editing the config file, so that it picks up the changes.
Hopefully this helps!
8458850–1122812–accelerator.cfg.default.txt (3.05 KB)
Good morning
Thanks for your reply. The AgentId in my config file was indeed empty. The documentation does not list it as something that needs to be set, it does not seem to be included in the default install.json that is created with the headless installer. Is this field usually populated by the installer, e.g. taking the hostname of the machine?
I can reach the dashboard now, but I see this in the logs:
{"level":"warn","ts":"2022-09-23T09:43:03.647Z","msg":"NO PROTOCOLS ENABLED","agent_id":"unity-accelerator1","agent_name":""}
netstat shows the dashboard port exposed now, but nothing more:
tcp 0 0 127.0.0.53:53 0.0.0.0:* LISTEN 3459/systemd-resolv
tcp 0 0 127.0.0.1:25 0.0.0.0:* LISTEN 413/master
tcp6 0 0 ::1:25 :::* LISTEN 413/master
tcp6 0 0 :::8443 :::* LISTEN 15657/unity-acceler
tcp6 0 0 :::22 :::* LISTEN 1/systemd```
[quote="SockPuppetFriend, post:2, topic: 894998, username:SockPuppetFriend"]
Then, if a lot of things are different between this default config and what you have, then I suggest reinstalling and choosing the most basic install instructions:
* *Existing Configuration Found*, choose *Replace*...
* *Protocol Configuration*, choose *Accelerate Asset Import Pipeline*; uncheck *Accelerate Unity Collaborate*; leave *Registration Key* blank
* *Security Configuration*: *Allow insecure access with NO encryption*
[/quote]
By the description I assume that you are referring to the UI driven installer. I did not mention this specifically, but this is a headless container I want to run the accelerator on.
Could you point me to which steps I need to take to configure *Accelerate Asset Import Pipeline* in a headless environment?
Edit: Found it. I could enable Asset Import Pipeline Version 2 caching via the Dashboard after logging in, the settings value is "ADBV2Enabled".
For reference:
Sensitive information replaced with [omitted]
8460059–1123178–install.json.txt (3.72 KB)
8460059–1123181–unity-accelerator.cfg.txt (3.18 KB)
I will look into this. We should have a working procedure for doing a headless install and this should be documented somewhere discoverable. Thanks for bringing the issue to my attention. I’ll do a little investigation and will share what I find here.
Hello again. I figured it out, it seems to be a minor bug.
The issue is that AgentID does not get populated by the headless installer when ADBV2 is not enabled. It does get populated correctly though when ADBV2 is enabled. However at this time, almost all users are using ADBV2, so this issue will not come up very often. I’ll file a bug for it. At some point we may remove Collab support completely and then ADBV2 should be enabled by default.