Reasons Kubernetes is so complex

Re:Reasons Kubernetes is so complex by cyberpunk on Thursday January 27, 2022 @11:33AM (#30102544)

VMware can do it by live migrating the vm, you will incur a short pause though, and the networking is a bit tricky to setup.. This of course doesn't happen during an unexpected downtime, it's a cold boot on another node in that case.

[ Parent | on HN ]

Re:Reasons Kubernetes is so complex by ted_dunning on Thursday January 27, 2022 @01:30PM (#30104223)

If you migrate a VM to a second hardware server then you have, by definition, a second server.

The question was how you reboot the posited singular hardware server with no downtime to any VMs running on it.

[ Parent | on HN ]

Re:Reasons Kubernetes is so complex by bpodgursky on Thursday January 27, 2022 @12:26PM (#30103332)

I have never seen this go smoothly on a production server; it's always WAY slower than expected (if you use any significant amount of memory) and something always gets f'd up wrt the network connectivity, broken caches, etc.

[ Parent | on HN ]

Re:Reasons Kubernetes is so complex by dekhn on Thursday January 27, 2022 @01:07PM (#30103869)

if I understand correctly, this is happening (nearly) transparently on GCP all the time. https://cloud.google.com/compute/docs/instances/live-migrati...

It involves copying the whole VM image over, and rewiring the network connections virtually on the fly.

I say (nearly) because it's not 100% transparent, but my understanding is that it works properly the vast majority of time.

[ Parent | on HN ]

Re:Reasons Kubernetes is so complex by bpodgursky on Thursday January 27, 2022 @02:33PM (#30105132)

Probably true but I'm also pretty sure Google isn't using VMware for live migrations under the good.

[ Parent | on HN ]

Re:Reasons Kubernetes is so complex by ryan_lane on Thursday January 27, 2022 @01:36PM (#30104301)

And what do you do when you need to restart the actual VM itself? You know, for when you need to patch the kernel and such...

[ Parent | on HN ]