minishift: minishift fails to restart if starting with service-catalog
General information
- Minishift version: minishift v1.10.0+10461c6
- OS: Mac OS X 10.13.2
Steps to reproduce
- Start a fresh minishift with servicecatalog: MINISHIFT_ENABLE_EXPERIMENTAL=y minishift start --openshift-version=v3.7.0 --memory 4GB --service-catalog --extra-clusterup-flags “–loglevel 1 --service-catalog”
–> minishift starts cleanly 2. Stop minishift 3. Start up minishift again with the same options MINISHIFT_ENABLE_EXPERIMENTAL=y minishift start --openshift-version=v3.7.0-rc.0 --memory 4GB --service-catalog --extra-clusterup-flags “–loglevel 1 --service-catalog”
Expected
Minishift starts up again cleanly
Actual
– Starting OpenShift container … I0112 14:14:17.445219 3806 helper.go:585] Copying OpenShift config to local directory /var/folders/bt/c4_bx2yx3h51f5zb96b79gfr0000gp/T/openshift-config522530176
Starting OpenShift using container ‘origin’ Waiting for API server to start listening OpenShift server started – Registering template service broker with service catalog … FAIL Error: cannot register the template service broker Caused By: Error: cannot create objects from template openshift-infra/template-service-broker-registration Caused By: Error: an error on the server (“Error: ‘dial tcp 172.30.1.2:443: getsockopt: no route to host’\nTrying to reach: ‘https://172.30.1.2:443/apis/servicecatalog.k8s.io/v1beta1/clusterservicebrokers’”) has prevented the request from succeeding (post clusterservicebrokers.servicecatalog.k8s.io) Error during ‘cluster up’ execution: Error starting the cluster.
Logs
You can start Minishift with minishift start --show-libmachine-logs -v5 to collect logs.
Please consider posting this on http://gist.github.com/ and post the link in the issue.
About this issue
- Original URL
- State: closed
- Created 6 years ago
- Comments: 38 (24 by maintainers)
@bparees, reusing. Minishift did create the config in the VM that reuses when it starts. If you see my (comment)[https://github.com/minishift/minishift/issues/1910#issuecomment-357641231], if you pass the flag 2 consecutive times, it fails. If on the second run, you no longer provide the flag, it runs successfully. This means that the catalog provisioning logic is not idempotent, or does not check if it has already been provisioned.