Base Command Manager / Bright Cluster Manager Release Notes

Release notes for Bright 8.1-13

== General ==

- New Features

* Upgrade docker to 17.03.2, Kubernetes is now nvidia-docker 2 ready
* Upgrade Singularity to 2.6.0
* Allow the exact DHCP interfaces to be set in the boot role
* Do not preselect by default the failover internal net in a Bright HA setup
* Allow prometheus timestamps to include [+-]XXXX timezone information

- Fixed Issues

* An issue with symlinks for cm-component-certificate tool
* Possible cmdaemon crash when Etcd is being uninstalled
* all2xml.py example doesn't export monitoring data producer arguments
* ldapserver entry missing from /etc/hosts on cloud director with direct connect
* Improved deployment of Kubernetes on HA setup
* Slurm configuration file update was not always triggered after setup
* An issue with the logging configuration for OpenStack Neutron and Cinder services

== cmdaemon ==

- Improvements

* An issue with workload management job time daylight handling

- Fixed Issues

* An issue with writing bonding options in the case of a bridge on top of the bond
* A new metrics collection script sample_smart to work around issues with the collection of metrics reported by smartctl tool

== node-installer ==

- Fixed Issues

* Allow to pass mount options for diskless layout

== cm-scale ==

- Fixed Issues

* In some cases, an issue with the automatic undrain of the nodes
* cm-scale may stop the extra node if it is not needed for one resource provider, but still needed for another