update floatingip status tracebacks

[Salvatore Orlando]

Logs are showing plenty of tracebacks like the following: http://logs.openstack.org/96/66796/15/check/check-tempest-dsvm-neutron-full/d940e56/logs/screen-q-svc.txt.gz?level=DEBUG#_2014-03-06_10_14_46_352

An operation (either update_floatingip_status or get_floating_ips) in update_floatingip_statuses (l3_rpc_base.py) triggers this traceback. Apparently this happens because if a floatingIP is removed, the transaction is aborted because of the exception.

Optimizing as suggested here: https://github.com/openstack/neutron/blob/master/neutron/db/l3_db.py#L680 might solve the issue.

[Oleg Bondarev (obondarev)]

ok I'll mark bug 1287049 as duplication of this one

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/79443
Committed: https://git.openstack.org/cgit/openstack/neutron/commit/?id=31f051a0349a40fe629a7bb5ce0ea683aa28a659
Submitter: Jenkins
Branch: master

commit 31f051a0349a40fe629a7bb5ce0ea683aa28a659
Author: Salvatore Orlando <email address hidden>
Date: Mon Mar 10 13:42:58 2014 -0700

    Optimize floating IP status update

    Hopefully also avoid tracebacks due to nested transactions
    being rolled back

    Change-Id: I1e74fb8d90de09b53d330f499b93073f19a5b9bf
    Closes-Bug: #1290549