Hoster v0.3

So many changes have been introduced since the v0.2b that it's really hard to list them all. So I'll keep the list short by describing only the major changes.

The list of major changes for v0.3

• HA Mode has been introduced. Now it's possible to run groups of (loosely coupled) HA nodes, and automatically failover the VMs in case there are any issues with a particular cluster node. HA depends on the REST API being available on each node, including the incoming http port. HTTPs hasn't been implemented yet, so please use WireGuard or Nebula to secure the traffic.
• REST API is being improved in every minor release. It's now picking up the configuration from restapi_config.json instead of the command line flags.
• CPU Threads is now supported as an optional key in the vm_config.json:

"cpu_threads": 2

• PCI Passthru is now supported as an optional key in the vm_config.json:

"passthru": [ "1/0/0", "1/0/1" ]

or

"passthru": [ "-1/0/0", "-1/0/1" ]

you can find out about the differences from the Issue tracker or our docs.

• New command has been introduced: hoster passthru list, which returns a similar output:

╭─────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────────╮
│                                                             Bhyve Passthru Devices                                                              │
├────────────┬─────────────┬────────────────┬──────────────┬────────────────────┬──────────────────────────────────────────┬──────────────────────┤
│ PPT Device │ PCI ID Raw  │      Type      │ Bhyve PCI ID │       Vendor       │                  Model                   │        Status        │
├────────────┼─────────────┼────────────────┼──────────────┼────────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ ppt0       │ pci0:1:0:0: │ display/VGA    │ 1/0/0        │ NVIDIA Corporation │ GP107GL Quadro P600                      │ In use by: test-vm-1 │
├────────────┼─────────────┼────────────────┼──────────────┼────────────────────┼──────────────────────────────────────────┼──────────────────────┤
│ ppt1       │ pci0:1:0:1: │ multimedia/HDA │ 1/0/1        │ NVIDIA Corporation │ GP107GL High Definition Audio Controller │ In use by: test-vm-1 │
╰────────────┴─────────────┴────────────────┴──────────────┴────────────────────┴──────────────────────────────────────────┴──────────────────────╯

• Added an initial support for FreeBSD Jails. You can now deploy, start and stop Jails using Hoster. Snapshot, replication and config reset will be coming in the next release.
• Added node_exporter_custom to improve the Prometheus integration

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.

Release v0.2

The list of all major changes:

• Added vm/info-all REST API endpoint
• Fixed broken (stale) VM uptime REST API response
• Changed VNC resolution to 800x600 to improve the noVNC performance
• Fixed host info wait group overflow
• Added the overall number of VMs on any given system to the /host/info API endpoint
• Added --force flag to vm stop command
• Fixed disk expand VM exists check
• Added vCPU to pCPU ratio to the host info table output
• Added --wait-time flag to vm start-all command

For the list of all changes, just check the Git log.

Test before upgrade

If you'd like to test this release first, download all 3 binaries into a separate folder, and use them locally, ie ./hoster vm list. P.S. don't forget to include your config files folder!

Upgrade

To start using this release, download and replace all 3 binaries manually, or use the self_update_service for the in-place update.

Release v0.1c

Changes made:

• Added support for Ubuntu 22.04
• Built and published Ubuntu 22.04 on the main public OS image server

Release v0.1b

List of changes in this release:

• Added the ability to show remotely available OS images

Release v0.1a

Changes in this release:

• Fixed CI reset option: there was a bug in the CI reset process, where it could not carry overt the old CPU cores and RAM values into the config generated on a new host. The workaround was to manually populate the missing values.
• Fixed version output string to present 0.1a
• Added self_update_service

Release v0.1

Trying to fix logo size