Skip to content

Improve / fix start/stop scripts#90

Open
johanib wants to merge 2 commits intomainfrom
feature/fix-issues
Open

Improve / fix start/stop scripts#90
johanib wants to merge 2 commits intomainfrom
feature/fix-issues

Conversation

@johanib
Copy link
Copy Markdown
Contributor

@johanib johanib commented Apr 30, 2026

Restart managegui in init.sh to reload Shibboleth SP metadata

After docker compose down -v, the Shibboleth metadata cache is gone.
When containers restart, managegui's Shibboleth SP tries to fetch IdP metadata from EngineBlock before init.sh has run migrations, causing a 503 and a 10-minute retry delay that breaks Manage login.

This change ensures a managegui restart after the Manage push so Shibboleth re-fetches the metadata while EngineBlock is fully initialized and ready.


Fix docker start & stop scripts

Prior to this change the test profile was not started by default. This is needed to run the chrome behat suite in EB, so we usually need this.

Prior to this change, the core/stop-dev-env.sh would not bring down the test profile. This caused orphaned containers, which produced issues related to the docker network not being able to be removed, which prevented a clean startup of the core environment.

Prior to this change, the docker compose -v flag was not able to be passed to the stop scripts, making it difficult to clean the slate.

johanib added 2 commits April 30, 2026 12:28
After `docker compose down -v`, the Shibboleth metadata cache is gone.
When containers restart, managegui's Shibboleth SP tries to fetch IdP metadata from EngineBlock before init.sh has run migrations, causing a 503 and a 10-minute retry delay that breaks Manage login.

This change ensures a managegui restart after the Manage push so Shibboleth re-fetches the metadata while EngineBlock is fully initialized and ready.
Prior to this change the `test` profile was not started by default. This is needed to run the chrome behat suite in EB, so we usually need this.

Prior to this change, the `core/stop-dev-env.sh` would not bring down the test profile. This caused orphaned containers, which produced issues related to the docker network not being able to be removed, which prevented a clean startup of the `core` environment.

Prior to this change, the docker compose `-v` flag was not able to be passed to the stop scripts, making it difficult to clean the slate.
@johanib johanib requested a review from kayjoosten April 30, 2026 11:25
@johanib johanib moved this from New to In Progress in PHP development Apr 30, 2026
@johanib johanib self-assigned this Apr 30, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

Status: In Progress

Development

Successfully merging this pull request may close these issues.

1 participant