Closed prateeksahu closed 3 years ago
hi @prateeksahu, we currently support only Ubuntu 18 for the host OS (albeit I hear that Ubuntu 20 should be fine but we haven't officially moved to it just yet).
Deploying in a VM should not be an issue, as long as you configure the VM to support KVM (i.e., expose VT-x to the VM).
Could you please try to reproduce the problem on a Ubuntu 18 host OS and check that the host VM has VT-x/KVM enabled?
Also, can you please list the exact commands that one can run to reproduce the problem?
Hey @ustiugov
Thanks, I found out the i had not enabled the VT-x on my VM. That seems to have solved the issue. We can close this bug.
One query though, upon clean exit of the vhive
program, any restart fails with port 3334/3333 as being busy. Any clue into this?
@prateeksahu glad to hear!
vHive executable listens on these ports but should have been freed after a graceful shut down... Nevertheless, you can run the following script for a complete cleanup:
./scripts/github_runner/clean_cri_runner.sh firecracker
Describe the bug Example functions deployment gives
CreateContainerError
in a single node cluster deployment.Setup Single node vHive deployment on VM (x86 machine). Host OS: Ubuntu 20.04, Guest OS: Ubuntu 18.04 Linux: 4.15
Expected behavior Deployer should be able to deploy the functions
Logs
firecracker-containerd logs:
Above logs repeat a lot. vhive logs:
Similar logs repeat for vhive too.
I understand that VM is currently not supported, but I would like to understand what is the cause for this. Running bare-metal on 20.04 also causes the setup script to break sometimes unexpectedly hence I gave VM a try.