-
Bug
-
Resolution: Not a Bug
-
Blocker
-
None
-
None
-
Incidents & Support
-
False
-
-
False
-
-
-
Critical
-
None
Description of problem:
The Submariner gateway nodes connection status is degraded.
E0717 02:37:32.159062 1 queue.go:146] "Unhandled Error" err="local -> broker for *v1.Endpoint: Failed to process object with key \"submariner-operator/di1001-submariner-cable-di1001-10-204-252-125\" using function (workqueue.ProcessFunc)(0x1923100): error distributing resource \"submariner-operator/di1001-submariner-cable-di1001-10-204-252-125\": error creating or updating resource: error retrieving \"di1001-submariner-cable-di1001-10-204-252-125\": Get \"https://api.cpaas-tc1002.gkee.p3.openshiftapps.com:443/apis/submariner.io/v1/namespaces/anz-onprem-gnet-set-broker/endpoints/di1001-submariner-cable-di1001-10-204-252-125\": read tcp 10.204.252.125:39876->10.54.254.190:80: read: connection reset by peer - error from a previous attempt: read tcp 10.204.252.125:34874->10.54.254.191:80: read: connection reset by peer" logger="UnhandledError"
- ACM Hub is on ROSA HCP. The managed clusters are baremetal with masters and ingress on VMware. The managed clusters themselves are on a flat network.
- Restarted lighthouse-agent, routeagent and gateway pods on both managed cluster but that did not help.
- It was configured with vxlan cabledriver, tried it with libreswan as well but getting same error.
- Also, added `forceUDPEncaps: true` but that did not make any difference.
Version-Release number of selected component (if applicable):
ROSA HCP 4.18.18
ACM 2.13.3
MCE 2.8.2
Submariner version: 0.20.1
Managed cluster OCP 4.16.34
How reproducible:
Always on customer environment
Steps to Reproduce:
- ...