Details
-
Bug
-
Status: Resolved (View Workflow)
-
High
-
Resolution: Fixed
-
29.0.7, 30.0.0
-
Security Level: Default (Default Security Scheme)
-
None
-
Horizon 22 - March 16 - 30
-
Backlog
Description
When starting the latest snapshot of Horizon 30 via Docker (i.e., opennms/horizon:bleeding, I get "null" on the graphs page.
To reproduce the problem, start a simple lab via Docker, for instance:
--- services: database: container_name: database hostname: database image: postgres:12 environment: - POSTGRES_HOST=database - POSTGRES_PORT=5432 - POSTGRES_USER=postgres - POSTGRES_PASSWORD=postgres ports: - 5432:5432 healthcheck: test: pg_isready -U postgres interval: 5s timeout: 1s retries: 10 opennms: container_name: opennms hostname: opennms image: opennms/horizon:bleeding depends_on: database: condition: service_healthy command: [ -s ] ports: - 8980:8980 - 8101:8101 environment: - POSTGRES_HOST=database - POSTGRES_PORT=5432 - POSTGRES_USER=postgres - POSTGRES_PASSWORD=postgres - JAVA_OPTS=-Xms4096m -Xmx4096m -XX:+UseG1GC -XX:+UseStringDeduplication healthcheck: test: curl -f -I http://localhost:8980/opennms/login.jsp interval: 30s timeout: 1s retries: 10
As you can see, I'm using the defaults (no changes in OpenNMS configuration besides access to the database).
By default, OpenNMS will add a node to monitor itself. You can add more nodes if you need them. From the node page, when you click on Resource Graphs, then either select a few and click Graph Selection or use Graph All, the URL you'll see on the Browser's navigation bar has "null" instead of "graph/results.htm" which is what OpenNMS must use.
http://localhost:8980/opennms/null?generatedId=94118d42-d552-3a3e-b541-b2a006ae5826&reports=all
If you manually update the URL, I can see the graphs appropriately rendered.
http://localhost:8980/opennms/graph/results.htm?generatedId=94118d42-d552-3a3e-b541-b2a006ae5826&reports=all
I'm having issues compiling OpenNMS from source, so I don't know if the problem is associated with the Docker deployment or not.
I even tried using a Private Browser Tab to discard potential cache-related problems, and the problem persists.
