HN Explorer

Ask HN: Simple cluster self-healing system?

2 points by 7532yahoogmail ↗ HN
I'm looking for a Linux based solution that, like the simplest uses of kubernetes, can run, monitor to insure N copies of task is running somewhere on a specified cluster on reachable hosts. I'm getting to kubernetes, but need something simpler first.

6 comments

[ 2.9 ms ] story [ 22.5 ms ] thread
hitsurume
If you're planning to run your own cluster, throw out the word simple. If you want simple, pay GCP or AWS for their solutions.
7532yahoogmail
To further clarify all I need is a system to:

- ask for task heartbeat

- make sure nodes in cluster pingable

- run N copies of task on reachable hosts

That's it.

ablekh
I would recommend you to consider the following "simpler-than-Kubernetes" containerized workload orchestrators: Swarm (https://github.com/docker/swarm) and somewhat more complex Nomad (https://github.com/hashicorp/nomad). Perhaps, even simpler orchestrators exist. If you're OK with having some vendor lock-in, you could also consider Amazon Elastic Container Service (ECS) or AWS Fargate or Azure Container Instances (ACI).