mrobson / atlantaistio Goto Github PK

Describe the bug
When accessing chat-client and chat-server URLs, an HTTP 503 response is returned from the istio-envoy sidecar instead.

Expected behavior
Return the application content and an HTTP 200 status code.

Steps to reproduce the bug

Attempting to access the chat-server URL as follows. For brevity, I've omitted the output of chat-client, but it returns the same HTTP 503 as chat-server.

# curl http://chat-server.app.example.com -kI
HTTP/1.1 503 Service Unavailable
date: Thu, 11 Apr 2019 19:56:18 GMT
server: istio-envoy
transfer-encoding: chunked
Set-Cookie: e223e97dadf2fc335d8e439cd4af9fe8=5f39c53105d0512caf9c53942801c373; path=/; HttpOnly

The following is the virtualservice as provided in the Git repo:

# oc get virtualservice -o yaml chat-server
apiVersion: networking.istio.io/v1alpha3
kind: VirtualService
metadata:
  creationTimestamp: 2019-04-11T14:49:25Z
  generation: 1
  name: chat-server
  namespace: chat
  resourceVersion: "2680043"
  selfLink: /apis/networking.istio.io/v1alpha3/namespaces/chat/virtualservices/chat-server
  uid: fff17e41-5c68-11e9-8139-021a4a0ab221
spec:
  gateways:
  - server-gateway
  hosts:
  - chat-server.app.example.com
  http:
  - match:
    - uri:
        prefix: /
    route:
    - destination:
        host: chat-server
        subset: v1

I have also tested the Bookinfo tutorial from the Service Mesh installation instructions, and it seems to work fine. I then swapped the VirtualService objects between bookinfo and chat-server and bookinfo fails while chat-server succeeds. This leads me to believe that there is something inherently incorrect with my setup and configurations relating to the chat-server/chat-client app pods themselves, not the gateway, virtualservice, destinationrules, or routes.

Version
I am using the Red Hat Service Mesh images, version 0.9.0 provided by registry.access.redhat.com/openshift-istio-tech-preview

sh-4.2# mixs version
version.BuildInfo{Version:"0.9.0-1", GitRevision:"7bcedaaa5365089ce59f50bebc840b569c6c8dd5eb", User:"redhat", Host:"redhat", GolangVersion:"go1.11.4", DockerHub:"brew-pulp-docker01.web.prod.ext.phx2.redhat.com:8888/openshift-istio-tech-preview", BuildStatus:"Clean", GitTag:"unknown"}

Environment
Lab cluster running on OpenShift 3.11 / ServiceMesh 0.9.0

I've deployed the chat-server service and it seems to be generating repeated errors in the logs:

Error ParserError: Protocol error, got "\u0015" as reply type byte. Please report this.
Error AbortError: Fatal error encountert. Command aborted. It might have been processed.
Error AbortError: Ready check failed: Fatal error encountert. Command aborted. It might have been processed.
Error ParserError: Protocol error, got "\u0015" as reply type byte. Please report this.
Error AbortError: Fatal error encountert. Command aborted. It might have been processed.
Error AbortError: Ready check failed: Fatal error encountert. Command aborted. It might have been processed.

This was an issue with both chat-server-v1 and chat-server-v2. Unfortunately, the disk space filled up so quickly that it overran my node's Docker storage within a couple of days and I had to shut down the pod.