What is npd-kdebug

node-problem-detector aims to make various node problems visible to the upstream layers in the cluster management stack. It is a daemon that runs on each node, detects node problems and reports them to apiserver. node-problem-detector can either run as a DaemonSet or run standalone. Now it is running as a Kubernetes Addon enabled by default in the GCE cluster.

In this example, we integrate the node-problem-detector with kdebug. After deploying kdebug with a specific check mode and node-problem-detector, kdebug will try to detect potential problems under the check mode. The results of the check will pass to node-problem-detector, and be reported through problem APIs of node-problem-detector.

• NodeCondition: Permanent problem that makes the node unavailable for pods should be reported as NodeCondition.
• Event: Temporary problem that has limited impact on pod but is informative should be reported as Event.

We call the integration of node-problem-detector(npd) and kdebug as npd-kdebug.

How to deploy npd-kdebug

We have already prepared a sample yaml file for you to help you deploy the intergration application of npd-kdebug with DNS check mode. You can run the following command to deploy the integrated daemon app to your kubernetes cluster.

kubectl apply -f ./node-problem-detector.yaml

What can npd-kdebug show you

Check the npd-kdebug is ready

• In kubernetes dashboard, you can click Daemon Sets in the side bar.
• If you see information like the following picture, it means that npd-kdebug is working on your cluster.

Check the problem detecting result of kdebug check

• Click Cluster > Nodes, and select a node.
• In the 'Conditions' tag, you can see a 'DNSProblem' type. It is a type of problems that detected by kdebug, and reported to node-problem-detector, as the node-problem-detector finally shows the Status and Messages.
  • If Status=False, it means there is no DNS problem.
  • If Status=True, it means npd-kdebug detected some DNS problems, and error messages show in Messages.

Customization

Besides DNSProblem check, you can integrate other kdebug check modes with npd. To customize different check modes npd-kdebug, you can follow the step-by-step tutorial in this section.

• Step 1: Copy the template yaml and open it.

• Step 2: Replace <YOUR_JSON_NAME> with a json file name you want at line 42. Recommend to include the check mode name you want to deploy. For example, kdebug-http.

• Step 3: Now you should edit your config json. This part of contents describe the parameters that how you run the kdebug, which is as the custom plugin of npd.

• Step 4: You should replace <YOUR_JSON_NAME> with the json file name you entered in Step 2.

• Step 5: In "conditions", you should fill the values of type, reason and message, which will be showed in dashboard if the mode check by kdebug passed.

• Step 6: You could define rules by edit the rules property. For the npd supporting two different types of check, you should define two rules of temporary and permanent types. Actually, you can define both rules by offering the same parameters of kdebug commands.

  You could replace <CONDITION_NAME> and reason as you want. For <CHECK_MODE> in args property, please replace it with the name of check mode('http' for HTTP check, e.g.). Due to you could only define certain conditions in Step 5, you should include the flag -c in args, or kdebug would execute all modes of check. The -f flag would make the output of kdebug check formatted, which will be showed as Message in dashboard. For more supported arguments of kdebug you can use, please refer to the help messages by running the following command.

  kdebug -h

  

• Step 7: Now you can use your customized yaml file to deploy your npd-kdebug by following How to deploy npd-kdebug