Speckle Deployment: Deploying a Server - manual setup produced some errors

[Terencema]

Hello dear Specklers, I have follow the Deploying a Server - manual setup on ubuntu server. When I run docker run the server, it show me an error saying

Thanks for your support

Best,
Terence

[jonathon]

Are you following the self deployment with VM manual? Deploying a Server - manual setup | Speckle Docs

[iainsproat]

Hi @Terencema

Welcome to Speckle’s community!

Can you please run docker system info and provide the value of the Architecture property?

This error typically occurs if the computer you are running on is Apple silicon (M1/M2), the architecture of Apple’s silicon is arm64 (or aarch64), which is not guaranteed to be compatible with the amd64 images we produce.

Iain

[iainsproat]

I tried to recreate your issue using an Ubuntu 22.04 x64 virtual machine (I used a Virtual Machine from DigitalOcean):

After installing [docker], I was able to run speckle-server version 2.15.0 as expected, and was unable to recreate this issue:

The architecture on the Ubuntu VM is reported by Docker as x86_64:

Iain

[Terencema]

I am using raspberry pi to install the ubuntu server. And this are my architecture:

[iainsproat]

Unfortunately aarch64 architecture is not currently supported.

[gjedlicska (Gergő Jedlicska)]

Hey @Terencema

we haven’t looked into rpi support yet, for the simple reason, that even under simple load, the full server stack can produce a considerable amount of load on the host system.

Keep in mind, when you are deploying a full server stack, you deploy the frontend, backend, fileimport, preview and webhook services. The RAM requirement of these services alone will stretch the capabilitie of most pi-s (except maybe the pi4 with 8GB ram).

While we do not officially support this (and most probably will not in the near future), IIRC there are no technical limitations, so if you really want to give this a go, you can, but you would need to rebuild the docker containers of the speckle services on the pi itself.
A partial example of how to do this with docker compose (my recommended route) can be found here https://github.com/specklesystems/speckle-server/blob/f2f3f9078681eef85bdff06a6ebe70b72468db7f/docker-compose-speckle.yml

BUT to be honest, you’d probably have a much better experience trying out our digital ocean 1-click app, which we officially support Speckle Server | DigitalOcean Marketplace 1-Click App