-
Bug
-
Resolution: Done
-
Medium
-
Beijing Release, Casablanca Release
-
None
Netconf mount does not persist in a 3-node cluster after deleting pod.
Install OOM with appc replica count = 3. After appc is fully initialized, add a netconf mount by issuing a PUT with something like the following:
PUT /config/network-topology:network-topology
{
"network-topology": {
"topology": [
{
"topology-id": "topology-netconf",
"node": [
,
{ "node-id": "58369c90-620a-4e7a-b358-a6e74e6266ee", "netconf-node-topology:keepalive-delay": 120, "netconf-node-topology:between-attempts-timeout-millis": 2000, "netconf-node-topology:host": "10.12.6.39", "netconf-node-topology:reconnect-on-changed-schema": false, "netconf-node-topology:password": "admin", "netconf-node-topology:username": "admin", "netconf-node-topology:sleep-factor": 1.5, "netconf-node-topology:port": 2831, "netconf-node-topology:tcp-only": false, "netconf-node-topology:connection-timeout-millis": 20000, "netconf-node-topology:max-connection-attempts": 0 } ]
}
]
}
}
Now, delete one appc pod with kubectl delete pod dev-appc-0 -n onap. Wait for the pod to get restarted by k8s.
Try retrieving the topology repeatedly by running the following in a loop:
curl -u admin:admin 'http://10.12.6.13:30230/restconf/config/network-topology:network-topology
The result is that a third of the time, the result returned does not contain the netconf mount. It appears that the mount is lost in the node that got recreated.