News for hackers. Stuff that matters.
Re:Four years after its release, Kubernetes has come a long way by zzzcpan on Wednesday June 06, 2018 @05:31PM (#17251283)
In fact, kubernetes won't solve high-availability for you. It is completely orthogonal to failover.

EDIT: No, really, k8s won't make a typical RDBMS suddenly able to run in three datacenters across three continents.

Re:Four years after its release, Kubernetes has come a long way by pests on Wednesday June 06, 2018 @06:52PM (#17251793)
But it can. You can develop a custom controller that does manage a typical RDBMS to be able to run in three datacenters across three continents.
Re:Four years after its release, Kubernetes has come a long way by pritambaral on Wednesday June 06, 2018 @09:41PM (#17252591)
> You can develop a custom controller that does manage a typical RDBMS to be able to run in three datacenters across three continents.

Honestly, at that point, what does k8s get you?

AIUI, Kubernetes is fine for stateless systems, but it is really no better on the storing-state question.

Re:Four years after its release, Kubernetes has come a long way by wmf on Wednesday June 06, 2018 @11:33PM (#17253044)
It's easier to install Stolon under K8s than by hand. :-)
Re:Four years after its release, Kubernetes has come a long way by pests on Thursday June 07, 2018 @01:27AM (#17253497)
Consistency.
Re:Four years after its release, Kubernetes has come a long way by zzzcpan on Thursday June 07, 2018 @11:36AM (#17256747)
It was a trick sentence, RDBMSs that can do that and remain useful enough for at least some applications don't exist.
Re:Four years after its release, Kubernetes has come a long way by shaklee3 on Wednesday June 06, 2018 @09:35PM (#17252570)
Huh? One of the core features is making sure you have N pods running at all times. If one fails, it starts another.
Re:Four years after its release, Kubernetes has come a long way by zzzcpan on Thursday June 07, 2018 @11:32AM (#17256703)
This feature won't even work for anything designed for fault tolerance. I believe it can only work for things that either rely on other properly fault tolerant services or that are completely stateless, idempotent, so they can be retried and eventually consistent (in which case the state still has to be handled by properly fault tolerant systems). Either way kubernetes cannot help here.