Closed ameier38 closed 5 years ago
I think this relates to the EVENTSTORE_EXT_IP_ADVERTISE_AS
env being set to statuspodIP
in the statefulset. If you remove this env, you will notice that you can access both eventstore-admin
and eventstore
service via port-forward.
We have already talked about related stuff over here. However, you mentioned if we don't set this env var, clustering would be broken.
Currently my team uses my fork with the removed env variable as we don't need to cluster eventstore yet but rely heavily on port-forward. I'd be very happy to see a solution here!
Yea I think you are right. I have been trying to wrap my head around the networking, specifically how the gossip works, but going to need help.
Are you able to connect to the database using port-forward? I can only seem to get the admin interface working if I remove the EVENTSTORE_EXT_IP_ADVERTISE_AS
as you did. When I was testing this a while back I thought I was able to get it working but now I am running into issues.
I try to connect using the following:
helm install -n eventstore .
kubectl port-forward svc/eventstore 2113
But it fails using discover://localhost:2113
as the uri. If I connect from inside the cluster using discover://eventstore:2113
then it works fine.
Here is a bit more info about advertise-as and when use it https://eventstore.org/docs/server/ports-and-networking/index.html?#advertise-as I'm also trying to fix that issue. It doesn't look right having both EXT_IP and EXT_IP_ADVERTISE_AS both assigned with status.podIP
@hnicke https://github.com/EventStore/EventStore.Charts/pull/30 should allow port-forwarding out of the box against a single node.
Is this a BUG REPORT or FEATURE REQUEST? (choose one): BUG REPORT
Version of Helm and Kubernetes: Helm v2.12.0 Kube v1.10.11
Which chart: eventstore
What happened: Installing chart and then running
kubectl port-forward svc/eventstore-admin 2113
works but when navigating tolocalhost:2113
the admin interface is not available and looks like it is receiving the wrong ip address.What you expected to happen: The admin interface would be available through port-forward
How to reproduce it (as minimally and precisely as possible):