In the MCP 2019.2.3 maintenance update, Mirantis introduces the following enhancements for Ceph:
To obtain the enhancements, follow the steps described in Apply maintenance updates.
TECHNICAL PREVIEW
Improved Ceph performance monitoring by implementing new Ceph prediction alerts. The new alerts include prediction of the IOPS consumption per OSD and pool, available RAM on the Ceph nodes, OSD disks responsiveness based on the write and read latency in a defined time range, as well as prediction whether a pool can consume all available capacity in a defined time range.
You can enable Ceph prediction only if you have previously enabled the Ceph Prometheus plugin as described in MCP Operations Guide: Enable the Ceph Prometheus plugin.
Improved the Add a Ceph OSD node and Ceph - replace failed OSD Jenkins pipeline jobs by integrating the upmap mechanism to fine control the placement group (PG) mapping, as well as by integrating the balancer plugin to provide for better control while adding new Ceph OSD nodes or replacing a failed Ceph OSD node.
To obtain the pipeline jobs improvements:
On the cluster level of the Reclass model, add the following class in
classes/cluster/CLUSTER_NAME/cicd/control/leader.yml
to add the
upmap-based pipeline job:
classes:
- system.jenkins.client.job.ceph.add-osd-upmap
(Optional) If you have used the Add a Ceph OSD node Jenkins pipeline job:
system.jenkins.client.job.ceph.add-osd
class from the
same file.Apply the jenkins.client
state on the cid01
node:
salt cid01\* state.sls jenkins.client
Verify that the Ceph - Add OSD pipeline job is available in the Jenkins web UI.