Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

As far as I recall it's not a VM. They run in "LX Branded Zones" which does require a Linux userland so that the binaries can find their libraries etc but Zones are more like "better cgroups than cgroups, a decade earlier" than VMs.


No, it's a VM, running a bhyve-based hypervisor, Propolis.[0] LX branded zones were/are great -- but for absolute fidelity one really needs VMs.

[0] https://github.com/oxidecomputer/propolis


Do you have a solution for running containers (Kubernetes, etc)? Are you spinning up a Linux VM to run the containers in there, doing VM per container, or something else?


Costumers can decide I would assume. Most likely you install you install some Kubernetes and then just have multible VMs distributed across the rack. And then run multible Pods in each node.

VM per container seems like a waist unless you need that extra isolation.


I wondered if there was any support for running containers built in - something like EKS/AKS/GKE/Cloud Run/etc - but looking at the docs it appears not.

I agree that VM per container can be wasteful - though something like Firecracker at least helps with start time.


From the podcast it seems that they want to deliver a minimal viable product. Their primary costumers already have a lot of their own higher level stack.

They might get into adding more higher level software eventually depending on what costumers want.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: