[zmq] Message sending shouldn't fail when redis is not available

[Oleksii Zamiatin]

In ZeroMQ driver we use Redis as a point of coordination where nodes/services can discover each other. When nodes discovery done and connections has being established we don't need Redis any more, until we want to add some more nodes/services/proxies. Following this logic we don't need Redis-HA (as we do with sentinel now, btw sentinel is experimental feature so using it in production is a risk).

Currently driver fails with error when Redis is not available in the system.

Expected: Each time driver couldn't update from Redis warning message is being logged and work continues with cached targets state and on existing connections.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/352579

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/352579
Committed: https://git.openstack.org/cgit/openstack/oslo.messaging/commit/?id=f61f0c1c1b037ac182c36ee8b98d2418beaa82a4
Submitter: Jenkins
Branch: master

commit f61f0c1c1b037ac182c36ee8b98d2418beaa82a4
Author: Oleksii Zamiatin <email address hidden>
Date: Mon Aug 8 15:27:00 2016 +0300

    [zmq] Redis unavailability is not critical

    Each time driver couldn't update from Redis
    warning message is being logged and work continues
    with cached targets state and on existing connections.

    Change-Id: I459532f1f60fab1d4c9926e02cdc3c3c6a162047
    Closes-Bug: #1610958

[Davanum Srinivas (DIMS) (dims-v)]

This issue was fixed in the openstack/oslo.messaging 5.9.0 release.