synchronization-testsuites

The main purpose of this repo is build the necessary tooling to collect necessary synchronization-related data logs from a running OpenShift cluster. This data will then be leveraged by different analysis tools to determine if the cluster is running within acceptable bounds synchronization-wise.```

The core approach taken is to strongly encourage and enforce separation of concerns between:

1. Declarative description of the cluster(s) under test
2. Configuration of a test (e.g. number of repetitions, acceptable thresholds, etc.)
3. Collectors - methods of collecting indicative information about the cluster
4. Checks - performed on collected values

Setup

1. Install Go
2. Install dependencies with go mod tidy

Development Extras

1. Install dev binaries: make install-tools. Ensure your $GOBIN is on your $PATH
2. yamllint
   1. Install yamllint with sudo yum install yamllint
   2. run with yamllint ./
3. golangci-lint
   1. Install golangci-lint
   2. run with make lint
4. license-eye
   1. Install license-eye with go install github.com/apache/skywalking-eyes/cmd/license-eye@latest
   2. run with license-eye header check or license-eye header fix
5. pre-commit
   1. on RHEL, pre-commit requires recompiling python to include optional sqlite modules:
      1. sudo yum install sqlite-devel
      2. See instructions here
   2. install pre-commit with pip3.10 install pre-commit
   3. configure your repository to run pre-commit hooks with pre-commit install
   4. manually run against all files with pre-commit run --all-files or against staged files with pre-commit run.
   5. Otherwise pre-commit will now run automatically when you make a new commit.

Usage

Building binary

Run the following command to build the binary used in the following commands:

go build

NOTE: use the --race flag when developing collectors.

Checking Enviroment

Run the following command (check help string for more details):

./collector-framework env verify --interface="<ptp interface>" --kubeconfig="${KUBECONFIG}"

Running Collectors

Run the following command (check help string for more details):

./collector-framework collect --interface="<ptp interface>" --kubeconfig="${KUBECONFIG}"

Fetching logs

The log subcommand has been removed. Instead we have implimented at collector which is enabled by default. If possible you should use a log aggregator. You can control the collectors running using the --collector flag.

Running tests

TODO: implement tests for all packages

To test the framework components run ginkgo pkg/<packagename>, for example to run the unit tests for the config package use ginkgo pkg/config

Contributing to the repo

See Adding a collector

To Do List

• unit tests for all of pkg/
• add more collectors
• better data persistance options