Hey!
I am new to FreeBSD and was captivated by its elegance pretty much from the first moment on.
Given its excellent hypervisor performance (I am talking about Bhyve of course), I am tempted to write a **lightweight** VM orchestrator for things such as code compilation on different platforms, automated tests, etc. This mainly consists of creating new VMs, running them, and sending a few console commands. The bhyve CLI tools seem amazing and intuitive, but of course for automation one would prefer to not parse plaintext. I was wondering if there is a way from e.g. C to run bhyve commands without posix_spawn() "hackery". I know libvmmapi exists, but that is way too low-level for my purposes. So what is the usual way of writing such an automation/"orchestrator"? Is it truly by parsing plaintext returned by spawning CLIs? Or am I missing something? Would there be demand for someone (possibly me?) refactoring bhyve into a library foremost and rewriting the CLI to use that library (looking at the source, this task is not too daunting due to the high quality modularization of the code)? Or is my thinking far off from that of a UNIX user?
Thanks and Cheers!
I am new to FreeBSD and was captivated by its elegance pretty much from the first moment on.
Given its excellent hypervisor performance (I am talking about Bhyve of course), I am tempted to write a **lightweight** VM orchestrator for things such as code compilation on different platforms, automated tests, etc. This mainly consists of creating new VMs, running them, and sending a few console commands. The bhyve CLI tools seem amazing and intuitive, but of course for automation one would prefer to not parse plaintext. I was wondering if there is a way from e.g. C to run bhyve commands without posix_spawn() "hackery". I know libvmmapi exists, but that is way too low-level for my purposes. So what is the usual way of writing such an automation/"orchestrator"? Is it truly by parsing plaintext returned by spawning CLIs? Or am I missing something? Would there be demand for someone (possibly me?) refactoring bhyve into a library foremost and rewriting the CLI to use that library (looking at the source, this task is not too daunting due to the high quality modularization of the code)? Or is my thinking far off from that of a UNIX user?
Thanks and Cheers!