How to determine why the precise-code-intel-worker is in CrashLoopBackOff status in Kubernetes deployment

This document will discuss one of the main reasons why the precise-code-intel-worker goes into a CrashLoopBackOff state in a Kubernetes. It does not attempt to solve all reasons why this state can happen.

This commonly happens when upgrading from a Sourcegraph version prior to 3.22 to a later version and failing to deploy the MinIO container. This is because in 3.21 -> 3.22 we removed the code intel bundle manager and replaced it with MinIO.

Symptoms

When running kubectl get pods you notice that the precise-code-intel-worker is in this state, example:

BASH
precise-code-intel-worker-9b69b5b59-jl6vd   0/1     CrashLoopBackOff   416        2d5h
precise-code-intel-worker-9b69b5b59-z7xx4   0/1     CrashLoopBackOff   415        2d5h

Steps to resolve

  1. Check what version of Sourcegraph you are on. If it is 3.22 or later, you will need to deploy MinIO because the precise-code-intel-worker depends on MinIO to function. If 3.4.2+, then minio is no longer used and instead sourcegraph/blobstore is used.

  2. Check what pods you have deployed and make sure MinIO is in the list.

    kubectl get pods -o wide

  3. If MinIO is not deployed, create a fork of the deploy-sourcegraph repository and make sure you deploy MinIO (or blobstore in 3.4.2+).

Further resources