collect_rabbitmq_stats.sh should write to a temp file prior to making stats available

[Benjamin Kaehne]

collect_rabbitmq_stats.sh should write to a temp file prior to making stats available.

While reading information from rabbit, the script should stage all writes to a temp file then make stats avail to any scripts which may read off this file.

The current issue is that any script which reads the stats file while it is being generated could potentially read an empty or half written file resulting in false errors to the consumer of the file.

[OpenStack Infra (hudson-openstack)]

Fix proposed to branch: master
Review: https://review.openstack.org/508849

[OpenStack Infra (hudson-openstack)]

Reviewed: https://review.openstack.org/508849
Committed: https://git.openstack.org/cgit/openstack/charm-rabbitmq-server/commit/?id=a8f7dea190a6e698d7851642fd16b67c5ed4f6fa
Submitter: Jenkins
Branch: master

commit a8f7dea190a6e698d7851642fd16b67c5ed4f6fa
Author: Liam Young <email address hidden>
Date: Mon Oct 2 11:00:14 2017 +0000

    Write stats to tmp file before updating data file

    This change updates the stats collection script so it writes to a
    temporary file before updating the main stats file to ensure that
    the update is an atomic operation.

    Change-Id: I68d4e3e723ed62fb5e90ba6bcf284c432ba9cd6b
    Closes-Bug: #1569143