Background : I'm "stuck" with and "old" infra, trying to configure a new nexus registry using swarm docker, behind traefik v1.7.
I'm trying to use "segments", as stated in the doc
docker service create \
--name nexus \
--label service-name=nexus \
--label namespace=nexus \
--container-label namespace=nexus \
--network tools-network \
--publish 8082:8081 \
--publish 5001:5000 \
--limit-cpu=1 \
--reserve-memory=2GB \
--limit-memory=3GB \
[...]
--label traefik.nexus_admin_prd_vdr.frontend.rule=Host:nexus.example.org \
--label traefik.nexus_admin_prd_vdr.port=8081 \
--label traefik.nexus_admin_prd_vdr.protocol=http \
--label traefik.nexus_admin_prd_vdr.loadbalancer.stickiness=true \
--label traefik.nexus_registry_prd_vdr.frontend.rule=Host:registry.example.org \
--label traefik.nexus_registry_prd_vdr.port=5000 \
--label traefik.nexus_registry_prd_vdr.protocol=http \
--label traefik.nexus_registry_prd_vdr.loadbalancer.stickiness=true \ --label traefik.docker.network=tools-network \
sonatype/nexus3:3.65.0
When I start this container, that makes a strange issue : it's "breaking" an other container (grafana), where some graph have no value for 2 weeks...
BUT.... If I correct the labels with :
--label traefik.frontend.rule=Host:nexus.eaxmple.org \
--label traefik.port=8081 \
--label traefik.protocol=http \
--label traefik.loadbalancer.stickiness=true \
--label traefik.registry.frontend.rule=Host:registry.example.org \
--label traefik.registry.port=5000 \
--label traefik.registry.protocol=http \
--label traefik.registry.loadbalancer.stickiness=true \
i.e. by removing one segment name, like "admin" (same behaviour if I remove "registry") :
-> All is working fine (Grafana container, admin frontend, and registry frontend)
-> but I don't see the "admin" frontend and backend in the Traefik console.
Edit : I just found out that the "segment" with no name defined is also not working in https (but OK in plain http)
Could you please help me fix my config ?