Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 8.0-19
== General ==
- New Features
* Mellanox 4.7 OFED Stack
== cmdaemon ==
- Fixed Issues
* Entity measurable cache is cleared before full stop, which in some cases can result in missing monitoring data after cmd restart
* Data overload in monitoring can cause crash in cache
* Monitoring for nodes can sometimes not be picked up on the head node
* In some cases, crash when cluster overview is called in parallel
* Jobs with negative exit code not saved to the DB
== cm-nvidia-docker ==
- Fixed Issues
* Apply a patch for NVIDIA docker to make it work with newer libglvnd-gles package
== cmsh ==
- Fixed Issues
* cmsh can hang when the table column size gets rescaled to 1 character
== slurm ==
- New Features
* Upgrade slurm to 18.08.8 and slurm19 to 19.05.2
* cmdaemon should be upgraded to the latest version prior to upgrading slurm to a more recent major version
* slurm-latest package is now deprecated and should be replaced manually with slurm or slurm19 as described in the Administrator Manual