DB lock wait timeout is possible for metaplugin

[Itsuro Oda]

There are some places that a target plugin's method is called in
'with context.session.begin(subtransaction=True)'.
This causes 'lock wait timeout' error potentially.
--- error is like:
OperationalError: (OperationalError) (1205, 'Lock wait timeout exceeded; try restarting transaction') ...snip
---

For example say ml2 is target plugin. Ml2's mechanism drivers separates precommit and postcommit
method so that 'lock wait timeout' error does not occur.
But it is meaningless if ml2 is used under metaplugin.

[yong sheng gong (gongysh)]

Is there someone still using metaplugin since we have ml2?

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/79534

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/79534
Committed: https://git.openstack.org/cgit/openstack/neutron/commit/?id=df315513efe62667220562dd2edb5401b15ab2ba
Submitter: Jenkins
Branch: master

commit df315513efe62667220562dd2edb5401b15ab2ba
Author: Itsuro Oda <email address hidden>
Date: Wed Feb 5 10:12:26 2014 +0900

    Call target plugin out of DB transaction in the Metaplugin

    Previously, there are cases that a target plugin is called
    within a DB transaction in the metaplugin. This possibly causes
    "DB lock timeout" error since a target plugin may take a long
    time (ex. communicate to a controller).

    Change-Id: Ie2971bfae1fecc97086b58cd943c321feb0aba04
    Closes-Bug: #1267291