Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 9.0-9
== General ==
- New Features
* Ceph support for RHEL8 (Octopus)
- Improvements
* Added CUDA 11.0 packages
- Fixed Issues
* Ensure LMod can load WLM modulefiles without specifying the full version
== cmdaemon ==
- Improvements
* Added an automatic retry for system info put, to overcome moments when the head node is too busy to process them
* Do not schedule image update for nodes which are down
* Show slurm accounting roles in cmsh show
* Clean up orphaned pexec trackers from the database to prevent slow cmd stop
- Fixed Issues
* An issue with the synchronization of kube role bindings
* An issue where /cm/node-installer-$distro-$arch exports are automatically disabled on an HA head node if there is no shared storage
* Disk information not shown for nvme drives
* Zombie monitoring keepalive tasks after running slurm steps
* TotalGPUUtilization metric value remains 0
* An issue with job monitoring of job arrays
* In some cases, an issue with bringing up a VLAN on top of a bond interface
* In some cases, an issue with collecting metrics for newly added Ceph OSD pools
== Bright View ==
- Fixed Issues
* Monitoring charts not refreshing in real time mode
== cmsh ==
- Improvements
* New islocked command in cmsh software image mode
* Tab completion for group member users
- Fixed Issues
* Crash in purgerevisionhistory cmsh command
== ml ==
- New Features
* Updated cm-chainer-* packages to v7.7.0
* Updated cm-theano-* packages to v1.0.5
* Updated cm-horovod-* packages to v0.19.5