Scaling out several vms failure when lb_policy attached

[EldonZhao]

I create a cluster and attached lb_policy to it. So when we scale out it with four nodes, it has an exception.

2016-07-17 23:40:34.293 4515 TRACE neutron.api.v2.resource raise loadbalancerv2.StateInvalid(id=id, state=status)

2016-07-17 23:40:34.293 4515 TRACE neutron.api.v2.resource StateInvalid: Invalid state PENDING_UPDATE of loadbalancer resource 0b2de551-782e-4744-9dce-eb55dda8e4e4

2016-07-17 23:40:34.293 4515 TRACE neutron.api.v2.resource

2016-07-17 23:40:34.294 4515 INFO neutron.wsgi [req-69b68ecb-13b0-4058-97a4-8860d4690065 ] 172.16.231.64 - - [17/Jul/2016 23:40:34] "POST /v2.0/lbaas/pools/34f325f7-5342-4bb8-b6d0-b30d1392e78c/members HTTP/1.1" 500 394 0.113535

We find that neutron doesn't support creating lb_member in parallel, and in senlin project, we ceate nodes in prallel.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/345195

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/345195
Committed: https://git.openstack.org/cgit/openstack/senlin/commit/?id=865f13dd0c9776836cf603fec463bb66b05066d4
Submitter: Jenkins
Branch: master

commit 865f13dd0c9776836cf603fec463bb66b05066d4
Author: yanyanhu <email address hidden>
Date: Thu Jul 21 02:46:46 2016 -0400

    Refactor lbaas driver

    This patch refactors lbaas driver to add workaround for
    concurrency issue happened during lb member adding/deleting.
    More work is needed to further optimize the design, e.g. make
    some important time settings configurable.

    Change-Id: Icc215dabe90a5c5e6dfb651dfe207d69b994f934
    Closes-Bug: #1604648

[OpenStack Infra (hudson-openstack)]

This issue was fixed in the openstack/senlin 2.0.0.0b3 development milestone.