See what happened¶
Find why a reef, a Nereid or an object is not Ready, starting from the CLI and going down to the object's events and the Nereid's log.
Before you start¶
- A cluster from Prepare the cluster, and
kubectlfor the last two steps.
Steps¶
1. See which reef has a problem¶
nereus status gives one row per reef and its Nereid:
nereus status
output
REEF NEREID VERSION STATUS AGE
authentication keycloak 0.1.0 Ready 2m
core built-in 0.1.0 Ready 4m
release flux 0.1.0 Ready 3m
secrets openbao 0.1.0 Ready 4m
tenancy - - Ready 4m
Registry localhost:5111
2. Look at the Nereid¶
nereus nereid get shows its image, digest and Mandates; a Nereid that is not Ready also says why (Unavailable, NotConformant).
nereus nereid get keycloak
output
Name: keycloak
Reefs: authentication
Image: localhost:5111/keycloak-controller:0.1.0
Digest: sha256:0f4675ab231b4e6d4a439d0c7d8e5b07011f35802b38592851542cf56666be93
Title: authentication-keycloak-controller
Description: The keycloak Nereid of Nereus, for the authentication reef
Status: Ready
Mandates: keycloak (default)
Age: 2m
3. Read the object's events¶
Each object's Ready condition says why in its message, and every change of a step is a Kubernetes Event on the object, kept when the Nereid restarts: kubectl describe shows them, under Events.
4. Read the Nereid's log¶
A Nereid logs JSON Lines, one per reconcile pass and its reason: kubectl logs deployment/keycloak --namespace nereus-system.
Check it worked¶
nereus doctor runs ten checks of the installation, pass or fail, and says the next step for one that fails:
nereus doctor
Next¶
- When the Nereid stops:
Unavailable, and how the commands come back. - Pass the conformance tests: the status every object must carry.
Learn more¶
- Spec 0003: the
Readycondition, the steps and their events.