Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 8.2-6
== cmdaemon ==
- New Features
* Allow bridges to be defined on top of VLAN interfaces
- Improvements
* Add cmdaemon table counts to cm-diagnose
- Fixed Issues
* OpenStack Heat will now use the public endpoint URL for generating URLs for triggering wait conditions. Previously it used the internal URL, which may not be reachable from within VMs.
* An issue with creating alias interface on bond or bridge on a SLES head node
* Switch port detection doesn't work for unknown OID
* Make ping by hostname thread safe
* Recognize Mellanox Technologies MT27800 as an IB interface in lspci
* Healthcheck 'schedulers' did not always work with slurm-latest
== node-installer ==
- Fixed Issues
* Allow node-installer to boot from tagged VLAN
== Bright View ==
- Improvements
* Run multiple PromQL queries in accounting and reporting
== cm-kubernetes-setup ==
- Fixed Issues
* The LabelSet for the Nvidia GPUs is automatically added during the setup
== cm-scale ==
- Improvements
* An issue with workload engine priorities not taken into account
== cmsh ==
- Fixed Issues
* Entering node execution filters mode directly in cmsh results in no entries
== cod ==
- Improvements
* Increased the speed of operations on multiple clusters by running them concurrently. New parameter "--max-threads" can be used to control how many operations are performed in parallel
== monitoring ==
- Improvements
* Add user name information to WAR metrics
- Fixed Issues
* Only filter on entity ownership in user based WAR reports
* Include node metrics for category in user-portal monitoring tree view
* Monitoring data is not cleaned up properly when Prometheus entities are removed
== user portal ==
- Improvements
* Run multiple PromQL queries in accounting and reporting
- Fixed Issues
* Recognize Mellanox Technologies MT27800 as an IB interface in lspci