Testgrid

Testgrid is a an automation testing platform for kURL. Testgrid installs kURL specs on a set of linux os versions and runs conformance tests.

Testgrid components and Architecture

  • Testgrid has three main components.
    • TGAPI: API is the main player that store and retrive the data to and from the database
    • Web: is the web interface for tests
    • TGrun: a daemon that polls the api for queued runs and starts new vms and tests.

test-gride-architecture

Run Testgrid Locally

Prerequests

   export GOOS=linux
   export GOARCH=amd64

Run Testgride using skaffold

  • Connect to the cluster
  • Run the following command from TESTGRID path
make install
  • Setup port-forwards

kubectl port-forward svc/tgapi 30110:3000 &
kubectl port-forward svc/testgrid-web 30881:30881

  • Now you are ready to do your first test.

  • From tgrun folder run the following command

./bin/tgrun queue --os-spec hack/os-spec.yaml --spec hack/test-spec.yaml --ref test-1 --api-token this-is-super-secret --api-endpoint http://localhost:30110

  • From the web service you should be able to see the pending test.

  • Now time to setup your runner by using terraform go indide the “deploy` folder and follow the steps from the readme file.

Run Testgrid on Okteto

  1. Change directories to the root of the project
  2. Run okteto pipeline deploy
  3. To “queue” a run ./bin/tgrun queue --os-spec hack/os-spec.yaml --spec hack/test-spec.yaml --ref ethan-1 --api-token this-is-super-secret --api-endpoint https://tgapi-${OKTETO_NAMESPACE}.okteto.repldev.com

Support for multi node clusters

  • Testgrid supports multi node clusters.
  • You can achieve this by putting the following lines in specs file, adjusting the number of nodes as desired

  numPrimaryNodes: 1
  numSecondaryNodes: 1
  • The previous example will make the cluster run with one primary node and one secondary node.

The process will be divided into three stages:

  1. Main process before the actual node run

  2. Initial primary node process

  3. Normal nodes process (other primary and secondary nodes)

Main process

  • The following chart explains what happens in the main process:

flowchart-mainprocess

InitPrimary node

The primary node can be in any one of the following statuses:

  1. running when the VMI start

  2. JoincommandStored when it does store the join command

  3. Success when the test run successfully

  4. Failed when any steps during the test fail

The following chart explains what happens in the initial primary node:

flowchart-initialprimarynode

Additional Nodes Flowchart

Additional nodes can be in the following status:

  1. Running when the VMI starts

  2. WaitingJoinCommand when waiting on the process to start

  3. Joined When the node is joined successfully to the cluster

  4. Success when the test runs successfully and the initial primary node finishes the test

  5. failed when it fails to run any step

The following chart explains what happens in the normal node (secondary and normal primary nodes):

flowchart-normalnode

GitHub

View Github