[bug] CSI volume endpoint returns 200 OK when payload is nonsensical

[jrasell]

Nomad version

Master commit 06f0f3a41563ac2cc557d8e68546ad6003b121ca

Operating system and Environment details

Reproduced on MacOS

Issue

It is possible for an operator to submit a register request to the CSI volume endpoint with a payload that is invalid and receive a 200 OK response. The register request successfully makes its ways via RPC and into raft. It misses any of Nomad's internal validation checks because the args.Volumes field is nil.

I believe inline with what I have seen in other agent endpoints, the decoded object should go through some basic validation before sending via RPC. Different endpoints approach validation in different ways:

• the operator endpoint uses a Validate function
• the job endpoint performs inline validation

Reproduction steps

• run Nomad in dev mode: nomad agent -dev
• using the payload below perform an API request: curl -v --request PUT --data @payload.json http://localhost:4646/v1/volume/csi/helloworld
• perform the curl request again and notice the X-Nomad-Index header integer has been incremented

example payload:

{"Foo": "bar"}

Nomad logs

I added a test log message to display the output of the rafApply function call which also shows the raft index being update on subsequent calls as detailed in the reproduction steps.

2020-05-20T11:50:58.625+0200 [DEBUG] nomad.csi_volume: ---------------------->: resp=<nil> index=10
2020-05-20T11:51:03.179+0200 [DEBUG] nomad.csi_volume: ---------------------->: resp=<nil> index=11

[tgross]

Thanks for reporting this, @jrasell. Easy enough to fix, but we should take a pass through the other endpoints to make sure we haven't dropped the ball on this elsewhere.

[jrasell]

Agreed @tgross and I have it on my list to do a quick pass when I have some more available time.

[github-actions]

I'm going to lock this issue because it has been closed for 120 days ⏳. This helps our maintainers find and focus on the active issues.
If you have found a problem that seems similar to this, please open a new issue and complete the issue template so we can capture all the details necessary to investigate further.