Outdated Version

You are viewing an older version of this section. View current production version.

Generate Cluster Reports

As of version 1.9.7, SingleStore DB Toolbox supports generating cluster reports via the Kube API for Kubernetes.

Info

SingleStore Tools must be installed on the same Linux host that has Kube API access to the cluster. This is a requirement as the Kube config is first used to make the connection, and then generate and pull the report.

SingleStore Tools may be installed as per Install SingleStore Tools.

There are two ways of running the collect-kube command to generate a cluster report:

  1. Run from anywhere using a kubeconfig file, which can be specified either from the KUBECONFIG environment variable or by the --config-file flag.

  2. Run from inside a Pod. Known as an in-cluster configuration, this method doesn’t require any configuration files. The service account should be configured properly to allow Kubernetes API calls, such as listing Pods.

    Note that the image for the Pod that runs sdb-report collect-kube must have the sdb-report binary available, such as via a memsql/tools image). Generally, this works just like kubectl does, with no installation required.

To generate a cluster report, run the following command.

The name of the cluster can be found by running kubectl get memsql.

sdb-report collect-kube -ns <name-space> --cluster-name <name-used-for-the-cluster>

Note: If the Operator logs are missing, run the following command.

kubectl label deployment <operator> app.kubernetes.io/component=operator

Refer to sdb-report collect-kube for more information.