Base Command Manager / Bright Cluster Manager Release Notes

Release notes for Bright 8.1-16

== General ==

- Improvements

* Added Intel 2019 compiler suite packages

== cmdaemon ==

- New Features

* Allow the switch ports to be ignored by setting the lowest port

- Fixed Issues

* Added extra check to prevent rsync running with empty exclude list
* An issue with terminating recently created spot instance
* In some cases, cmdaemon crash when repeatedly powering on a AWS spot instance
* Remove old entries from billing history
* Nodes not detected as down because the failBeforeDown values in the database can be incorrect
* An issue with lite nodes staying up after the node is rebooted when the service is not started
* Proc stat sampler is not automatically reinitialized when individual CPU sampling is turned on
* In some cases, network errors while accessing AWS APIs can cause cmdaemon termination
* AWS static IP not being released if AWS instance creation has failed
* An issue with viewing and assigning AWS static IPs
* LSF healthcheck reporting failed on some specific correct node states

== node-installer ==

- Fixed Issues

* An issue were on Ubuntu, IB interfaces were not brought up properly, which could among other result in an empty resolv.conf file

== cm-lite-daemon ==

- Improvements

* Add scripts to install required pip modules

- Fixed Issues

* An issue which causes ssl hostname checking when specifying the host as an IP
* In some cases cm-lite-cmdaemon doesn't detect IP / MAC properly

== cm-openstack-setup ==

- Fixed Issues

* Improved retry mechanism for running 'keystone-manage db_sync' during OpenStack deployment
* An issue with nova migration/resize due to missing SSH related configuration. The fix is part of openstack-nova-migration rpm installed on the hypervisors

== cod ==

- New Features

* Added ability to specify OpenStack credentials on the CLI. Added ability to provide not only domain IDs, but also Names. Improved password input prompt. Improved auth error message (easier to debug auth error).

- Improvements

* Added the ability for the cod client to send usage reports and runtime stats (disabled by default)
* It is now possible do delete a shelved cluster, without first needing to start (unshelve) it.
* Added an option to use a lightweight Bright cluster license activation server in COD setups without Internet connectivity

== monitoring ==

- Improvements

* lua environment for monitoring filtering

- Fixed Issues

* Optimize addition of new metrics
* An issue with monitoring after passive takes over