This message was deleted.
# harvester
a
This message was deleted.
d
Harvester is not packaged to be installed on top of an existing Linux distribution. It is built on an OS which is integrated into the package that it installs with. The specific OS package is derived from the "Elemental Toolkit" (https://rancher.github.io/elemental-toolkit/) which provides an underlying simple immutable Linux OS environment with the Harvester functionality layered on top of that. In principle, one could put all of the Harvester functionality aside from the Elemental Toolkit on top of some other Linux OS. It is, after all, all open source, so you could separate those things if you had the time, resources, and energy to do so. However, it's not advisable. Some parts of Harvester (such as etcd) are highly performance-sensitive. If you run Harvester in VMs or on a system with other unknown workload, there's a reasonable chance components of Harvester will perform poorly or even malfunction due to timing issues. So, if you want Harvester to perform reasonably and not malfunction, you really should run it on bare metal in the manner that is is packaged for distribution currently.
I should say that Harvester is designed to be a self-contained appliance and not to be an application co-existing on an OS with other applications.