Skip to content

Common issues

This article documents known issues that can occur during or after installation, along with their resolution procedure.

ArangoDB provisioning failure during PlatformInstallation

Affects: v0.0.1
Fix scheduled: Next release

Symptom

The PlatformInstallation command fails with the following error sequence in the logs:

{"module":"PlatformInstallation","job_provisioner":"ArangoDB","event":"Creating users","level":"info"}
{"module":"PlatformInstallation","job_provisioner":"KubernetesClient","event":"STDERR [kubectl ... exec -i arangodb-crdn-... -c server -- arangosh]: command terminated with exit code 1","level":"info"}
{"module":"PlatformInstallation","job_provisioner":"ArangoDB","event":"arangosh command returned an error","level":"info"}
{"module":"PlatformInstallation","job_provisioner":"ArangoDB","event":"Failed to execute arangosh command on coordinator","level":"info"}
{"module":"PlatformInstallation","job_provisioner":"Installer","event":"Exception was raised in step ArangoDB","level":"info"}
{"module":"PlatformInstallation","job_provisioner":"Installer","event":"One or more step failed, can't go further","level":"info"}
{"component":"platform-installer","exit_code":42,"event":"Job container exited with non-zero code","level":"error"}

The command exits with:

╭─────────────────────────────────── Error ────────────────────────────────────╮
│ install failed after 1 attempt(s)                                            │
╰──────────────────────────────────────────────────────────────────────────────╯

Cause

The ArangoDB operator occasionally fails to fully initialize before the platform installer attempts to provision users. This is a timing issue that occurs intermittently on first installation.

Resolution

Step 1: Wait for the operator to self-recover.

The ArangoDB operator recovers automatically in most cases. Wait 1 hour, then check pod status:

kubectl get pods -A | grep arango

If all ArangoDB pods show Running, proceed to retry the installation. If one or more pods are still not running after 1 hour, continue with step 2.

Step 2: Manually clean up the ArangoDB state.

Run the debug platform installer to open an interactive shell:

./run-shc.sh exec DebugPlatformInstallation
kubectl exec -it -n sekoia-system jobs/shpi-debug-pause-platform-installer-job -- bash

Inside the shell, run the teardown command. This command is expected to fail after a moment:

./platform-installer exec ArangoDB -m teardown

Exit the shell (exit or Ctrl+D), then delete the ArangoDB operator resources:

kubectl delete deployments.apps -n support arangodb-exporter
kubectl delete deployments.apps -n support arango-deployment-operator
kubectl delete secret -n support arangodb-operator-api-jwt arangodb-operator-api-jwt-key arangodb-operator-dashboard
kubectl delete cm -n support arangodb-operator-feature-config-map

Step 3: Retry the installation.

./run-shc.sh exec PlatformInstallation

If the installation fails again at the same ArangoDB step, repeat step 2 and step 3.