To troubleshoot the problem:
Verify the hypervisor at <hypervisor_ip>:8085/Snh_ItfReq
to identify
whether the configuration is present. For example:
http://10.92.249.119:8085/Snh_ItfReq?name=
If the configuration is missing, it means that the hypervisor did not receive it from the OpenContrail controller node.
Verify api-server
at <ip>:8082/virtual-machine-interface/<uuid>
to identify whether routing_instance_refs
are present. For example:
curl -u admin:secret123 http://localhost:8095/virtual-machine-interface/39f | \
python -m json.tool > /tmp/vmis
If routing_instance_refs
are missing:
Verify that contrail-schema
is running on at least one
OpenContrail controller node:
contrail-status
Verify that VMI is present in ifmap
on all OpenStack controller
nodes:
ifmap-view localhost 8443 visual visual | grep <vmi-name>
If the VMI is missing in ifmap
:
Verify that RabbitMQ is clustered successfully and all API servers are connected to it:
rabbitmqctl cluster_status
Verify that contrail-api
introspects the port at 8084
for
rest/db/messagebus/ifmap
.
To verify why contrail-schema
does not establish the link between
VMI and RI, inspect /var/log/contrail/schema.err
.
Inspect the OpenContrail logs:
contrail-logs --object-type config --object-values
contrail-logs --object-type config --object-id <id_from_above_command>
Inspect the latest log from OpenContrail to stdout
and other logs using
the contrail-logs command.
Verify the connection status, for example:
http://<config-node-where-schema-active>:8084/Snh_SandeshUVECacheReq?x=NodeStatus