* Don't bind 'user' role to test user
Signed-off-by: Dmytro Nochevnov <dnochevn@redhat.com>
* Remove role 'user' from keycloak
Signed-off-by: Dmytro Nochevnov <dnochevn@redhat.com>
Enable single-level DNS domains (*.domain.tld) in SINGLE_PORT mode to simplify HTTPS setup.
Use whole folder for traefik conf to inject SSL certificates and custom configuration.
CHE_SINGLEPORT_WILDCARD__DOMAIN_IPLESS to che.env.
Signed-off-by: Hanno Kolvenbach <kolvenbach@silexica.com>
Move openshift/k8s deployment files from dockerfiles/init folder
to deploy folder with a better structure because they are not
related to init dockerfile.
Add readme files in each deployment target folder with links to
Che docs where the process of deployment is described.
Remove delivery of deployment files by init image which is
useless.
Remove empty docs files from the repo and init image.
Fix some minor typos and trailing spaces.
Signed-off-by: Oleksandr Garagatyi <ogaragat@redhat.com>
* Experimenting with deployment config based approach to Che multi user on OCP
* Service pod
* Remove s2i
* Remove s2i
* EOF
* Cleanup
* Cleanup
* Cleanup
* Cleanup
* Make Keycloak util image configurable
* Update webdriver.sh
* Increase postgres and keycloak PVC for OpenShift to be 1gb
For compatibility and portability between OSO, OSD, and OCP this needs to be 1Gi as anything under that is not allowed in Online environments. Having it set to 100Mi results in errors and the deployment fails.
Cleanup and Roll Update not working due to PR #8158 exiting if existing project is not found. Moved command checking above project creation/check so that these commands can execute before scripts exits when it finds that project with the same name exists. (#8357)
Signed-off-by: James Drummond <james@devcomb.com>
Allow Openshift deploy_che.sh script to detect if project deletion is completed and new project successfully created.
Signed-off-by: James Drummond <james@devcomb.com>
The livenessProbe's failureThreshold is set to low. This causes the livenessProbe having to short a period of probing before it fails and removes the container. The failureThreshold is set to 4, the initialDelaySeconds is set to 5 and the periodSeconds is 5. This means 5s+4x5s=25s before the container fails. This patch set failureThreshold higher to 11 allowing for a more reasonable overall probe time of 60s (5s+11x5s=60s) before fails and removes the container.
Signed-off-by: James Drummond <james@devcomb.com>
Oleksandr Andriienko