Mongo Replicaset (kubernetes)

Kubernetes Mongo ReplicaSet

MongoDB ReplicaSet is a method to run a MongoDB database with high availability and redundancy by distributing the data across multiple MongoDB instances.

Design Decisions

1. Helm Deployment: The module uses Helm for deploying MongoDB ReplicaSet in Kubernetes. This leverages the Bitnami MongoDB Helm chart.
2. Immutable Infrastructure: Requests and limits are set to provide predictable scheduling and a "guaranteed class" of workloads.
3. Persistent Storage: Configured with persistent storage to ensure data durability.
4. Authentication and Security: Generates secure passwords for the root and application user. It also includes a replica set key for enabling replication.
5. Replica Configuration: Flexible configuration to set up the desired number of replicas.
6. Scalability and Upgrades: Easy to scale and upgrade with Helm chart versions and Kubernetes capabilities.

Runbook

MongoDB Connection Refused

If MongoDB is not accepting connections, check if the service and pods are running.

Check the status of MongoDB pods:

kubectl get pods -n <namespace> -l app.kubernetes.io/instance=<release-name>

If pods are not running, describe the pod to get more details:

kubectl describe pod <pod-name> -n <namespace>

MongoDB Authentication Issues

If you are facing authentication problems, ensure the credentials are correct.

Connect to MongoDB using the root user:

kubectl run -i --tty --rm mongo-client --image=mongo --restart=Never -- mongo --host <mongo-host> -u root -p <root-password> --authenticationDatabase admin

Replica Set Not Initialized

If the replica set is not properly initialized, check the replica set status.

Connect to MongoDB shell:

kubectl run -i --tty --rm mongo-client --image=mongo --restart=Never -- mongo --host <mongo-host> -u root -p <root-password> --authenticationDatabase admin

Check the replica set status:

rs.status()

If it shows an error, initialize the replica set:

rs.initiate()

Persistent Volume Issues

If the data persistence is not working correctly, check the status of Persistent Volume Claims (PVCs).

List PVCs in the namespace:

kubectl get pvc -n <namespace>

Describe the specific PVC:

kubectl describe pvc <pvc-name> -n <namespace>

MongoDB Performance Issues

If MongoDB performance is not optimal, check the resource usage of the pods.

Check CPU and memory usage:

kubectl top pods -n <namespace>

Additionally, log into the MongoDB shell and check slow queries:

db.system.profile.find({ millis: { $gt: 100 } })

This will display queries taking longer than 100 milliseconds. Adjust the threshold as needed.

Scaling MongoDB ReplicaSet

If you need to scale the number of replicas:

Edit the Helm values to change replicaCount and upgrade the release:

helm upgrade <release-name> bitnami/mongodb -n <namespace> --set replicaCount=<new-count>

Ensure the new pods are running and up to date:

kubectl get pods -n <namespace> -l app.kubernetes.io/instance=<release-name>