Nomad Tools Discussion - Githubissues

hashicorp / nomad

Nomad is an easy-to-use, flexible, and performant workload orchestrator that can deploy a mix of microservice, batch, containerized, and non-containerized applications. Nomad is easy to operate and scale and has native Consul and Vault integrations.

https://www.nomadproject.io/

Other

14.86k stars 1.95k forks source link

Nomad Tools Discussion #8816

Open jazzyfresh opened 4 years ago

jazzyfresh commented 4 years ago

As the Nomad community grows, the diversity of cluster computing use cases also grows. There is a greater need to have a swiss army knife of Nomad tools for understanding & interacting with the Nomad cluster.

This is a discussion ticket along that line, anyone should feel free to chime in with a potential need and/or use case. Then I will use the ticket description as a meta-ticket tracker for Nomad tools that we want to build. Whether these are Nomad CLI/API built-ins or Nomad ecosystem tools is up for discussion.

Some examples of tools/improvements:

pre-flight checker https://github.com/hashicorp/nomad/issues/8790
open tracing with Jaeger/Zipkin https://github.com/hashicorp/nomad/issues/4511

jazzyfresh commented 4 years ago

One thing I would like is a CLI that works very smoothly with bash scripting (eg, it emits Nomad alloc ids/task pids/etc for quick demos/debugging)

picatz commented 4 years ago

I have two, very similar ideas at the moment:

Something like this would be great to have: https://github.com/digitalocean/clusterlint
I've also been playing around with a related idea, using InSpec to test Nomad deployments: https://github.com/picatz/inspec-nomad/blob/master/controls/nomad.rb Moreover, InSpec integrates with Packer!

ketzacoatl commented 4 years ago

In no particular order:

"When I deploy this new cluster, help me navigate the vault/consul/nomad ACL tarpit"
"When I deploy this new cluster, make sure the following 'initial' (service/system/batch/whatever) jobs are running"
"When a node joins the cluster, run X or Y system jobs"
"Gracefully remove that node from the (consul/nomad) cluster (ineligible, drain, leave) and terminate/halt it" (see https://github.com/hashicorp/nomad/issues/2052)
"Operators can gracefully transition leadership" (see https://github.com/hashicorp/nomad/issues/7376).
"Give me a shell in the same context as this task alloc"
"Help me figure out why this task is failing or not running"

Thank you for the opportunity to share feedback on this!