Hi guys. I’m trying to set up a somewhat general-purpose dev environment for Sushi, and while I have managed to get it up and running in Ubuntu Studio, I’d like to try a more stripped back approach, maybe without a GUI/desktop environment that I can run in Docker or with Vagrant/Virtualbox etc. It looks like some of the required KXStudio tools can’t be controlled from the command line, though. Just wondering if you guys have managed to get anything like this up and running internally? Thanks.
For anyone looking at this, I’ve started putting together a Vagrant box running Ubuntu 20.04, with Ubuntu Studio installled over the top, and with sushi and all its dependencies installed, including gRPC etc. This would mean you simply clone a repo from github and run ‘vagrant up’, and the image will download and automatically start running in Virtualbox, and you’ll have a dir from the repo mounted on the system ready to start writing your code.
This dockerfile might help you out: https://github.com/ElliottLandsborough/sequensual/blob/main/Dockerfile
You are a god amongst men.
Pardon for my Docker noobness, but will the included Jack daemon actually work with audio when running it under macOS & Windows? In that case, that will be really very nice for those that don’t have a native Linux environment!
We finally just released a major update of Elk:
Check it out and let us know if it breaks your docker setup or if you need help with the updated dependencies.
It doesn’t work directly on Mac OS, but I’m actually running Mac OS, with a Linux VM in Parallels, and running Docker in the Linux VM, and it works flawlessly. I tried it with Virtualbox and had some issues because of VB’s buggy implementation for CoreAudio in newer versions of Mac OS, but Parallels has no issues at all. Basically spin up an Ubuntu VM in parallels, pull the repo and run. I ran the MDA-JX10 example using VMPK in my Linux VM, passing midi through to Sushi on the docker container and it worked well.
AMAZING. Thanks, can’t wait to try it out.
great to know.
Virtualbox’s virtual audio soundcard is notoriously bad for these use cases (and IIRC it doesn’t have input), a good workaround is to use a USB class-compliant soundcard and redirect that to the VM. In that way, you are also able to get almost decent latencies out of the VM.
Yeah, that’s how I’m working with Parallels, but I didn’t know Virtualbox could do that. Cheers for the tip!