Open Stack

Hi all

I've done the following steps while creating load-balancer and other entities under it.


1.      Create load-balancer.

2.      Created listener under load-balancer created in step 1

3.      Created pool under listener created in step 2

Here pool was created successfully but my load-balancer stuck in PENDING_UPDATE state in both Neutron & Octavia databases.

When I tried to delete pool I got error saying that pool can not be deleted because respective load-balancer is in immutable state. Ok fine.

After some time ( after around 1 hr) load-balancer in Neutron DB had become  ACTIVE state , but in Octavia DB still in PENDING_UPDATE state.

See below

$ neutron lbaas-loadbalancer-list

+--------------------------------------+----------+---------------+---------------------+----------+

| id                                   | name     | vip_address   | provisioning_status | provider |

+--------------------------------------+----------+---------------+---------------------+----------+

| 5f9d1b93-7997-44b5-8db5-53c312f0628c | test_lb1 | 50.50.50.3    | ACTIVE              | octavia  |

+--------------------------------------+----------+---------------+---------------------+----------+

mysql> select * from load_balancer;

+-----------+--------------------------------------+----------+-------------+---------------------+------------------+---------+----------+

| tenant_id | id                                   | name     | description | provisioning_status | operating_status | enabled | topology |

+-----------+--------------------------------------+----------+-------------+---------------------+------------------+---------+----------+

| NULL      | 5f9d1b93-7997-44b5-8db5-53c312f0628c | test_lb1 |             | PENDING_UPDATE      | ONLINE           |       1 | NULL     |

+-----------+--------------------------------------+----------+-------------+---------------------+------------------+---------+----------+

When I tried deleting the pool in such state , I got successfully deletion message as show below.

$ neutron lbaas-pool-delete 74cdfee8-dd49-4085-9830-d57933204c98

Deleted pool: 74cdfee8-dd49-4085-9830-d57933204c98

But pool is not really get deleted. See below

$ neutron lbaas-pool-list

+--------------------------------------+-------+----------+----------------+

| id                                   | name  | protocol | admin_state_up |

+--------------------------------------+-------+----------+----------------+

| 74cdfee8-dd49-4085-9830-d57933204c98 | pool  | HTTP     | True           |

+--------------------------------------+-------+----------+----------------+

This is a bug right ?  This bug a raised as load-balancer-state is not in synch in Neutron DB and Octavia DB.
This in synch is coming because load-balancer state becomes ACTIVE from PENDING_UPDATE automatically. Is this expected behavior ?
How can we make load-balancer state should be in synch in both the DBs.


Also I saw below log messages in Octavia-api.log

2016-10-04 05:57:28.865 10744 INFO octavia.api.v1.controllers.pool [req-cb15f3c9-8746-4079-945f-96cf112a26e2 - - - - -] Pool 74cdfee8-dd49-4085-9830-d57933204c98 cannot be deleted because the Load Balancer is in an immutable state
2016-10-04 05:57:28.896 10744 DEBUG wsme.api [req-cb15f3c9-8746-4079-945f-96cf112a26e2 - - - - -] Client-side error: Load Balancer 5f9d1b93-7997-44b5-8db5-53c312f0628c is immutable and cannot be updated. format_exception /opt/stack/venv/octavia-20160826T092820Z/lib/python2.7/site-packages/wsme/api.py:221


Thanks

Ravi
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openstack.org/pipermail/openstack-dev/attachments/20161028/0c62d9e2/attachment-0001.html>