Table of Contents |
---|
Working doc
https://docs.google.com/document/d/18Ow4R3tB9GIvdcFO7WmUvjb0a-sT6h0zSCEnfHsPz58/edit?usp=sharing
Everything below this line is currently a WIP taking place in the link above
---------------------------------------------------------------------------
Status
Current state: Under Discussion
...
To make it easy to run Cassandra on Kubernetes. If an operator is like a robot in your datacenter running your Cassandra cluster, what does that operator need: a) To make informed decisions. b) Exposed from node to cluster to take actions
Audience
- Devops
- Developers
- Ops
Goals
Operator Capability Level
(Taken from https://operatorframework.io/operator-capabilities/)
- Lower the impedance between Kubernetes and Cassandra operations
- Achieve Level 3 Operator compliance, plus Horizontal Scaling from Level 5.
- Provide pathway to Level4
- Listed on https://operatorhub.io/
What does Level 1, 2 and 3 look like in the Cassandra world?
E.g.
- L1: How do we support all the configuration options etc?
- L2: Patch management and pausing broken upgrades / running mixed versions etc.
Level 1 operator for Apache Cassandra
Level 2 operator for Apache Cassandra
Level 3 operator for Apache Cassandra
Non-Goals
- Remove the need for any Cassandra administration. (Not Level 5)
- Provide a serverless facade for Cassandra
- Official Docker images??
Proposed Changes
Describe the new thing you want to do in appropriate detail. This may be fairly extensive and have large subsections of its own. Or it may be a few sentences. Use judgement based on the scope of the change.A new repository as a sub-project for Apache Cassandra specifically for a Kubernetes Operator
New or Changed Public Interfaces
Briefly list any new interfaces that will be introduced as part of this proposal or any existing interfaces that will be removed or changed. The purpose of this section is to concisely call out the public contract that will come along with this feature.
A public interface is any change to the following:
- native protocol (and CQL)
- gossip and the messaging service
- pluggable components (SPIs) like authorisation, triggers, ..?
- commitlog, hintlog, cache files
- sstables components
- configuration
- jmx mbeans (including metrics)
- monitoring
- client tool classes
- command line tools and arguments
- operational routines
- - Better support for K8s service discovery mechanisms
- configuration - Probably something related to k8s service discovery e.g. a k8s seed provider maybe?
- The operator will define a Kubernetes Custom Resource (CRD), this will be the primary API for interacting with the operator.Anything else that will likely break existing users in some way when they upgrade
Compatibility, Deprecation, and Migration Plan
- Target version: 3.x and above
- What impact (if any) will there be on existing users?
- If we are changing behavior how will we phase out the older behavior?
- If we need special migration tools, describe them here.
- When will we remove the existing behavior?
Test Plan
Describe in few sentences how the CIP will be tested. We are mostly interested in system tests (since unit-tests are specific to implementation details). How will we know that the implementation works as expected? How will we know nothing broke?
Rejected Alternatives
...
- None
- The operator will need to support minor and major version upgrades of Apache Cassandra
Test Plan
- Used as a part of CI/CD for Apache Cassandra project
- dtest
- harry
- fallout
- TBD - Acceptance framework for k9s Operators. (Need guidance on that criteria)
Rejected Alternatives
- HELM charts