This change is required to fix different issues
related to order in which plugin tasks are launched
on controller nodes
Closes-Bug: #1558580
Change-Id: I8d26ca82e70ef80d509203366457779d6a3291a8
value of master name should be the same as in
ceilometer redis backend url
Co-Authored-By: alexey-mr <Alexey.Morlang@gmail.com>
Change-Id: Ic9f6399950597db399718b4e607c6586db5647e3
Closes-Bug: #1553229
In this patch next features have been implemented:
1. Set a notification workload_partitioning to True, for turning on
notification agents partitioning
2. Change alarm evaluator resource configuration in pacemaker to run
it on each controller node
Change-Id: Id74094d64b6c24aa4c4f4ff5d9a5cd2610425c60
In details:
* Adapt Redis puppet module for plugin usage:
- update tests
- improvements for sentinel installation
- uncomment CentOS params for plugin usage on next Fuel releases
* Prepare main puppet script for:
- Redis and sentinel installation
- Work with ceilometer-agent-central service as pacemaker resource
(remove old one and configure new with clone set)
- Configure redis under pacemaker
- Add ocf scripts for redis-server
* Add tasks for plugin
* Add metadata for plugin
* Add Redis packages (2.8 version) and their dependencies
* Use existing ceilometer network role for redis
Partially implements: blueprint ceilometer-central-agent-ha
Change-Id: I77c75e05b28f3ff4690e89034e345cc9c64af5e5
Add upstream Redis puppet module which will be used
as base module for Ceilometer-Redis plugin:
https://github.com/fsalum/puppet-redis
Partially implements: blueprint ceilometer-central-agent-ha
Change-Id: Icbd879623790bf96e54b6dad7435594f107cb650
Andreas Jaeger