Release v0.2b

Release notes

• Implemented vCPU to pCPU ratio. Now it will show up in the REST API host info and in the hoster host info table. Will add more docs about what it is later on down the road. This was a request from one of the users.
• Added backup VMs count to the REST API output. This makes it easier to implement WebUI representation of backup VMs.
• Added CPU info to the REST API and JSON outputs. WebUI will show this information on the host overview page.
• Vastly improved the log output for our REST API.
• Improved hoster vm stop vm-name --force flag. It allows to clean up the leftover VM resources in case of unpredictable circumstances.
• Implemented hoster vm stop-all --force flag. It's very useful on test nodes, where you want to kill all VMs at once.
• Added current host to VM info output and fixed parent host output. This will allow to improve the HA configurations and better represent the data in WebUI.
• Added hoster init check to all CLI commands. Now if you didn't run the hoster init after the node reboot, hoster utility will refuse to work until you do so.
• Fixed hard coded name servers search list. This configuration was present in the Cloud Init.
• Fixed hoster vm replicate bug (unpredictable behaviour if the VM didn't have any snapshots).
• Switched the default storage driver from virtio-blk to nvme.
• Vast VM Supervisor improvements, and big improvements to it's log output.

Admin notes

• VM Supervisor had a couple of critical flaws, that allowed undefined behaviour in case of VM reboot, which are now all fixed. The recommendation is to reboot your nodes that are running old (pre v0.2b) VM Supervisor binary, and then execute self_update_service or update all binaries manually, before you execute hoster init.
• Before you can reboot all of your nodes, the workaround is to download+replace just the hoster binary, and use hoster vm stop vm-name --force in case you VM is stuck.