LBaaS v2 healthmonitor wrong status detection

[Sergey Yezhkov]

Summary:
After enabling health monitor loadbalancer on any request returns
HTTP/1.0 503 Service Unavailable

I have loadbalancer with vip ip 10.123.21.15. HTTP listener, pool and member with IP 10.123.21.12.

I check status of web-server by:
curl -I -X GET http://10.123.21.15/owncloud/status.php
...
HTTP/1.1 200 OK

But when I add healthmonitor:
neutron lbaas-healthmonitor-create \
  --delay 5 \
  --max-retries 2 \
  --timeout 10 \
  --type HTTP \
  --url-path /owncloud/status.php \
  --pool owncloud-app-lb-http-pool

neutron lbaas-healthmonitor-show
+----------------+------------------------------------------------+
| Field | Value |
+----------------+------------------------------------------------+
| admin_state_up | True |
| delay | 5 |
| expected_codes | 200 |
| http_method | GET |
| id | cf3cc795-ab1f-44c7-a521-799281e1ff64 |
| max_retries | 2 |
| name | |
| pools | {"id": "edcd43a2-41ad-4dd7-809d-10d3e45a08a7"} |
| tenant_id | b5d8bbe7742540c2b9b2e1b324ea854e |
| timeout | 10 |
| type | HTTP |
| url_path | /owncloud/status.php |
+----------------+------------------------------------------------+

I expect:
curl -I -X GET http://10.123.21.15/owncloud/status.php
...
HTTP/1.1 200 OK

But result:
curl -I -X GET http://10.123.21.15/owncloud/status.php
...
HTTP/1.0 503 Service Unavailable

Direct request to member:
curl -I -X GET http://10.123.21.12/owncloud/status.php
...
HTTP/1.1 200 OK

In neutron logs have no ERROR.

Some detail about configuration:

I have 3 controllers. Installed by Fuel with l3 population and DVR enabled.
lbaas_agent.ini
interface_driver=openvswitch

neutron lbaas-loadbalancer-status owncloud-app-lb
{
    "loadbalancer": {
        "name": "owncloud-app-lb",
        "provisioning_status": "ACTIVE",
        "listeners": [
            {
                "name": "owncloud-app-lb-http",
                "provisioning_status": "ACTIVE",
                "pools": [
                    {
                        "name": "owncloud-app-lb-http-pool",
                        "provisioning_status": "ACTIVE",
                        "healthmonitor": {
                            "provisioning_status": "ACTIVE",
                            "type": "HTTP",
                            "id": "cf3cc795-ab1f-44c7-a521-799281e1ff64",
                            "name": ""
                        },
                        "members": [
                            {
                                "name": "",
                                "provisioning_status": "ACTIVE",
                                "address": "10.123.21.12",
                                "protocol_port": 80,
                                "id": "8a588ed1-8818-44b2-80df-90debee59720",
                                "operating_status": "ONLINE"
                            }
                        ],
                        "id": "edcd43a2-41ad-4dd7-809d-10d3e45a08a7",
                        "operating_status": "ONLINE"
                    }
                ],
                "l7policies": [],
                "id": "7521308a-15d1-4898-87c8-8f1ed4330b6c",
                "operating_status": "ONLINE"
            }
        ],
        "pools": [
            {
                "name": "owncloud-app-lb-http-pool",
                "provisioning_status": "ACTIVE",
                "healthmonitor": {
                    "provisioning_status": "ACTIVE",
                    "type": "HTTP",
                    "id": "cf3cc795-ab1f-44c7-a521-799281e1ff64",
                    "name": ""
                },
                "members": [
                    {
                        "name": "",
                        "provisioning_status": "ACTIVE",
                        "address": "10.123.21.12",
                        "protocol_port": 80,
                        "id": "8a588ed1-8818-44b2-80df-90debee59720",
                        "operating_status": "ONLINE"
                    }
                ],
                "id": "edcd43a2-41ad-4dd7-809d-10d3e45a08a7",
                "operating_status": "ONLINE"
            }
        ],
        "id": "67a9602e-4bcd-4d1c-a41c-7af20ded0300",
        "operating_status": "ONLINE"
    }

[Michael Johnson (johnsom)]

Hi Sergey,

What LBaaS driver and OpenStack release are you using?

[Sergey Yezhkov (yezhkov)]

I use OpenStack Mitaka installed by Fuel 9.0 and lbaas-plugin to install LBaaS.

I`m not sure how to find witch LBaaS driver I use.

In neutron.conf on controller I have
service_plugins=neutron.services.l3_router.l3_router_plugin.L3RouterPlugin,neutron.services.metering.metering_plugin.MeteringPlugin,qos,neutron_lbaas.services.loadbalancer.plugin.LoadBalancerPluginv2

[service_providers]
service_provider=LOADBALANCERV2:Haproxy:neutron_lbaas.drivers.haproxy.plugin_driver.HaproxyOnHostPluginDriver:default

Is it?

[Michael Johnson (johnsom)]

Ok, moving back to the neutron-lbaas bug queue. Sergey is using the HaproxyOnHostPluginDriver and not the octavia driver.

[Michael Johnson (johnsom)]

Are you still having this issue? I cannot reproduce it on my devstack.

If you can reproduce this, can you provide the commands you used to setup the load balancer (all of the steps), the output of neutron net-list, the output of neutron subnet-list, and the output of "sudo ip netns"?

[Sergey Yezhkov (yezhkov)]

I can't reproduce this bug anymore. But I update openstack since this bug happened.

At this moment on controller node I have this version of lbaasv2 agent
# apt list | grep neutron-lbaasv2
 neutron-lbaasv2-agent/mos9.0-updates,now 2:8.2.0-2~u14.04+mos2 all [installed]

Unfortunately I didn't log agent version at moment when bug happened.

[Launchpad Janitor (janitor)]

[Expired for octavia because there has been no activity for 60 days.]