Skip to content

Troubleshooting for Connector

Below are resolution strategies for the alerts

Topic Description
Alert - Connector Resource loading Error/Connector Permission loading Error Resolution steps and root cause when connector Resource loading Error/Connector Permission loading Error.
Alert - Connector Resource/Permission Loader Threads Shutdown Resolution steps and root cause analysis for Resource and Permission Loader thread shutdown alerts.
Alert - JDBC Error Rate Resolution steps and root cause analysis when connector encounters high JDBC query failure rates.
Alert - Event Processing Alerts Resolution steps and root cause analysis for High Event Processing Time and High Event Processing Lag alerts.
Alert - Changelog Task Queue Alert Resolution steps and root cause analysis when changelog processor queue count exceeds the threshold.
Alert - SDK API Error Rate High Resolution steps and root cause analysis when SDK API error rate exceeds the threshold.
Alert - RocksDB Total Memory Usage Resolution steps and root cause analysis when RocksDB total memory usage exceeds 1GB threshold.
PolicySync Changelog Slow-Task Diagnostics Opt-in, off-by-default per-phase slow-task WARN and automatic thread dump to localize a slow or stuck changelog.

Cleanup Connector RocksDB PVC in Existing Deployment

After restarting the connector following the PVC cleanup, the connector initializes as a fresh instance. All resources, permissions, and principals are loaded again, and Ranger policies are reapplied.

Perform the following steps to cleanup the Connector PVC.

DEPLOYMENT_ENV_NAME

You can get your deployment by running the command

Bash
1
2
cd ~/privacera/privacera-manager
grep DEPLOYMENT_ENV_NAME config/vars.privacera.yml
And set it as an environment variable
Bash
1
export DEPLOYMENT_ENV_NAME=<your-deployment-name>

  1. List all the deployments in the namespace.

    Bash
    1
    kubectl get deployments -n $DEPLOYMENT_ENV_NAME

  2. Identify the respective connector deployment and run the following command to scale down the connector deployment to 0.

    Bash
    1
    kubectl scale deployment <connector-deployment-name> --replicas=0 -n $DEPLOYMENT_ENV_NAME

  3. List all the PVC in the namespace

    Bash
    1
    kubectl get pvc -n $DEPLOYMENT_ENV_NAME

  4. Identify the respective connector PVC name. Run the following command to delete the connector PVC.

    Bash
    1
    kubectl delete pvc <connector-pvc-name> -n $DEPLOYMENT_ENV_NAME