Troubleshooting
Running into an issue while using Gloo Portal? Try the following troubleshooting topics to find why things might not behave as expected. If you are still stuck, try posting on the Solo.io Slack.
General steps before debugging
- Make sure that you run Gloo Portal in a supported environment:
- Gloo Edge Enterprise version 1.5.0-beta9 or later
- Kubernetes 1.16 or later
- Make sure that you access Gloo Portal from a supported browser.
- Review your infrastructure provider for any errors, incidents, or outages that might impact your clusters, load balancers, and related infrastructure. Common configuration problems happen in the following resources:
- Firewalls
- Application load balancer (ALB) settings
- Expired or malformatted certificates, tokens, and secrets
Gloo Edge gateways
Gloo Portal creates Gloo Edge VirtualServices in the same namespaces as the Gloo Portal Environment and Portal resources. Make sure you configured a
Gloo Edge Gateway resource
in such a way that the generated virtual services get applied to it.
The default Gloo Edge gateways match all virtual services in all the namespaces that Gloo Edge watches (which by default are all the namespaces).
Earlier versions of Gloo Portal CRDs
Early versions of Gloo Portal used to be installed as part of Gloo Edge Enterprise.
Before installing a new version of Gloo Portal, make sure your cluster does not have the older version of the Gloo Portal custom resource definitions (CRDs).
The older CRDs might have the same name as some newer ones, in which case Helm does not overwrite the CRDs. In turn, the older CRDs with a newer Gloo Portal installation might lead to errors similar to the following:
Error: no matches for kind "Portal" in version "portal.gloo.solo.io/v1beta1"
To remove these old CRDs, run the following command:
kubectl get crd | grep "devportal.solo.io" | awk '{print $1}' | xargs kubectl delete crd