Troubleshooting

A collection of guides, procedures, and utilities for troubleshooting CrateDB clusters.

You will learn how to apply self-service measures supporting you when observing problems with your CrateDB database cluster. If you need help from others, feel free to reach out to our community or commercial support channels any time.

General Recommendations

Relevant suggestions will support you to optimally gather and convey information about your database cluster, which helps others to get a good orientation about its technical details like cluster size, or data volumes. Before contacting support, you optimally have this information ready.

We recommend to use those guidelines as a checklist in all situations, in order to help our support engineers to discover any sort of problem with your CrateDB clusters much faster than needing to ask individual questions.

Note: Additional communication channels apply for users of CrateDB Cloud, or when running CrateDB on a commercial contract.

System Tables

Many details about the status of your cluster is included within CrateDB’s system tables. The sys schema includes synthetic read-only tables which can be queried to get real-time information about the cluster’s metadata, like information about nodes and shards.

Troubleshooting with system tables

How to use CrateDB’s system tables to investigate the database cluster status, and debug issues.

Diagnostics with System Tables

Diagnostic Utilities

Instructions about how to use relevant utility and diagnosis programs.

About CFR

The CrateDB Flight Recorder (CFR) collects information from CrateDB’s system tables, and bundles it into an archive file ready to share with our support engineers.

CrateDB Flight Recorder (CFR)
About jcmd

The jcmd utility is the traditional application to inquire diagnostics information from software running on the JVM. It also includes the Java Flight Recorder (JFR).

Using jcmd with CrateDB
About crate-node

A utility program to interact with a CrateDB cluster for conducting infrastructure operations. For example:

  • Control the master node election process.

  • Repurpose nodes: Detach and move between clusters.

  • Clean up stale node data.

The crate-node command
About crash

Troubleshooting the CLI program crash.

https://cratedb.com/docs/crate/crash/en/latest/troubleshooting.html

Note

You can find a lot of troubleshooting guides that explain how to perform diagnostics on Java applications.

Because CrateDB is written in Java, any of those tools can be used to troubleshoot CrateDB instances. Above, we are focusing on canonical and traditional utilities, for example jcmd.