Bug 1461624 - [free-int] Unable to access exposed service on cluster
Summary: [free-int] Unable to access exposed service on cluster
Keywords:
Status: CLOSED CURRENTRELEASE
Alias: None
Product: OpenShift Online
Classification: Red Hat
Component: Routing
Version: 3.x
Hardware: Unspecified
OS: Unspecified
high
high
Target Milestone: ---
: 3.x
Assignee: Clayton Coleman
QA Contact: zhaozhanqi
URL:
Whiteboard:
Depends On:
Blocks:
TreeView+ depends on / blocked
 
Reported: 2017-06-15 02:05 UTC by Justin Pierce
Modified: 2017-11-09 18:48 UTC (History)
10 users (show)

Fixed In Version:
Doc Type: If docs needed, set a value
Doc Text:
Clone Of:
Environment:
Last Closed: 2017-11-09 18:48:15 UTC
Target Upstream Version:
Embargoed:


Attachments (Terms of Use)


Links
System ID Private Priority Status Summary Last Updated
Origin (Github) 14818 0 None None None 2017-06-22 15:37:59 UTC

Description Justin Pierce 2017-06-15 02:05:42 UTC
Description of problem:
After creating a trivial app, I am unable to access it via its exposed route.

Version-Release number of selected component (if applicable):
oc v3.6.106
kubernetes v1.6.1+5115d708d7
features: Basic-Auth GSSAPI Kerberos SPNEGO

Server https://internal.api.free-int.openshift.com:443
openshift v3.6.106
kubernetes v1.6.1+5115d708d7


How reproducible:
100%

Steps to Reproduce:
1.    oc delete --ignore-not-found project aos-cd-smoke-test-proj
2.    oc new-project aos-cd-smoke-test-proj
3.    oc new-app --image-stream=ruby --code=https://github.com/openshift/ruby-hello-world
4.    oc logs -f bc/ruby-hello-world
5.    oc expose svc/ruby-hello-world
6.    oc get routes
7. Access the route with a browser

Actual results:
Hitting the URL (e.g. http://ruby-hello-world-aos-cd-smoke-test-proj.d800.free-int.openshiftapps.com/) reports:
-----------------------
Application is not available

The application is currently not serving requests at this endpoint. It may not have been started or is still starting.
...
-----------------------


Expected results:
The same process on another cluster with a slightly older oc level displays a "Welcome to an OpenShift v3 Demo App!" welcome message.

Additional info:

Comment 4 Ben Bennett 2017-06-15 15:55:40 UTC
jupierce and I hopped on the box to troubleshoot, and since we see reencrypt routes with no certs, and no matching cert written to disk... I suspect the culprit is in https://github.com/openshift/origin/pull/13752 somewhere.

It would help to have the output from 'oc get dc router -o yaml' as an attachment to the case.  Thanks.

Comment 7 Ben Bennett 2017-06-19 17:25:40 UTC
Justin, can you please add the router log 'oc log router-...' to the bug.

Comment 10 zhaliu 2017-06-20 10:11:13 UTC
The routes are available now except for reencrypt route.

free int:
openshift v3.6.116
kubernetes v1.6.1+5115d708d7

Comment 12 Ben Bennett 2017-06-21 18:32:37 UTC
Related:
 https://bugzilla.redhat.com/show_bug.cgi?id=1461378
 https://bugzilla.redhat.com/show_bug.cgi?id=1462709

You need https://github.com/openshift/origin/pull/13752 to make reencrypt routes work with your template.  Without that the router is crashing when someone makes a reencrypt route without a destiantion cert.

Comment 13 Clayton Coleman 2017-06-21 19:41:54 UTC
Go ahead and assign this to me, I'm going to make the oapi/v1 API backwards compatible in this case.

Comment 14 Ben Bennett 2017-06-22 15:42:21 UTC
*** Bug 1461378 has been marked as a duplicate of this bug. ***

Comment 15 Eric Paris 2017-06-27 15:32:47 UTC
https://github.com/openshift/origin/pull/14818

Comment 16 zhaliu 2017-07-03 02:42:06 UTC
@Clayton Coleman @Justin Pierce @Ben Bennett @Eric Paris @zzhao

We try to reproduce the problem in the devenv with free int router image , but the problem did not occur after creating and deleting the re-encrypt route with or without the destination ca cert.So could you possibly provide some help to reproduce this problem so that we could have a criterion to judge whether the bug was fixed?

The steps:
1. Launch the devenv with free int router image:
# ansible-playbook -i hosts devenv-launch.yml -e DEVENV=free
Modify the router image in dc:
registry.ops.openshift.com/openshift3/ose-haproxy-router:v3.6.116

2. Create a pod , a svc, and edge, passthrough, re-encrypt routes with or without destination ca cert:
# oc create -f https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/routing/caddy-docker.json
# oc create -f https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/routing/edge/service_unsecure.json
# oc create -f https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/routing/reencrypt/service_secure.json

# oc create route reencrypt reenroute--service=service_secure

# oc create route passthrough passroute --service=service_secure

# oc create route edge edgeroute--service=unservice_secure

# wget https://raw.githubusercontent.com/openshift-qe/v3-testfiles/master/routing/reencrypt/route_reencrypt_dest.ca

# oc create route reencrypt reencroute --service=service_secure --dest-ca-cert=route_reencrypt_dest.ca

# oc delete route reenroute reencroute

3. After the re-encrypt routes are created and deleted for many times, the errors like comment 9 are not found in the logs of router , but if we delete the destination ca cert in the router pod, the errors will be seen but the router can be running normally if accessing some other routes ,e.g: passthrough or edge.

Comment 17 zhaozhanqi 2017-07-04 06:26:04 UTC
since this bug did not found for free-int. so verified this bug


Note You need to log in before you can comment on or make changes to this bug.