-
Notifications
You must be signed in to change notification settings - Fork 5
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Attempting to run agent as pod fails #4
Comments
@erik-hallros curious if you ever got this working? I'm trying to deploy an agent inside the same cluster as well |
@jimmy-ungerman i had some success to get it running further by writing a custom middleware.sock that intercepts the kasm_agent path and instead fetches the container based on label instead of kasm_agent name, which works. But the more i dig into this the more i feel like i had the wrong expectations on how the agent worked, the way the agent application is written it seems to be meant to have only one agent on each host. Because what the agent image appears to be doing is to manage the host through the docker runtime api, it creates additional containers on the parent and mounts different directories from the host on each container instance. My assumption to begin with was that the agent image on dockerhub was a standalone agent container. |
@jimmy-ungerman I've made a container version that works but i doubt its best practices or remotely stable, done nothing in networking yet so theres stuff left to do to make it functional. Feel free to check it out or contribute |
K8s is currently only supported for the infrastructure of Kasm, not for the end-user desktop containers. For that we do support kubevirt auto-scaling, which uses full stack VMs for agents. There are people experimenting with DinD type setup for running agents in K8s. Ultimately we really need to create a K8s native agent that can create the end-user pods directly, rather than relying on DinD, but we are not there yet. |
@mmcclaskey Makes sense, I saw the blog post on friday that explicitly called out that agents aren't currently supported. Ideally, we can get a kubernetes agent that just spins up a new pod with the selected desktop container from a user that they can use |
When attempting to run a kasm agent as a kubernetes pod it seems to fail when executing the python application.
Is the container name hardcoded to kasm_agent in the python code?
To make this work up to this point i had to change the node runtime to docker and copy / replicate the behaviour of the docker agent created in the single host setup version by analyzing docker inspect output.
Pod Log
docker ps
The container name is not kasm_agent when running as pod with rancher, and underscore is not allowed in pod name when configuring the yaml file for kubectl.
agent.app.config.yaml
kasmagent.yaml
The text was updated successfully, but these errors were encountered: