Base Command Manager / Bright Cluster Manager Release Notes

Release notes for Bright 8.2-10

== General ==

- New Features

* Support for Jupyter Lab

== cmdaemon ==

- Improvements

* Improved check if the rsync exclude file was written correctly
* More descriptive messages on home directory create/move conflicts

- Fixed Issues

* In some cases, old active takes control back after "cmha makeactive" and powers off newly created active head node
* Jobs with negative exit code not saved to the DB
* snmp can time out for some community read strings, but not others
* An issue with job cgroup cache, which can lead to increased cmdaemon memory usage when running many jobs
* An issue with setting the hostname in /etc/sysconfig/network on nodes booting from disk
* Edge node can be falsely listed as installer failed / unreachable the first time they boot
* An issue with cmha nodestatus
* An issue with availability set support (needed for properly supporting VM types quipped with IB) for Azure
* An issue with saving the job information on cmd stop, which could leave monitoring data without a job and cause orphaned DB entries
* NFS export settings not properly imported from XML dump
* Tagged VLAN on top of BOOTIF can break assignment
* Make sure existing JupyterHub installations are updated correctly
* In some cases, clock shifts can prevent configuration files from being written
* In rare cases cmdaemon can end in 100% CPU usage
* Improved RPC timeouts

== cm-kubernetes-setup ==

- New Features

* Support for 3rd party / external Docker, new --skip-docker command-line flag to disable the Docker deployment

== cmsub ==

- Fixed Issues

* An issue with downloading stderr if it is different than stdout

== cod ==

- New Features

* New 'cm-cod-os cluster console' command

- Fixed Issues

* An issue with "cm-cod-os image install" and "cm-cod-os cluster tag"