Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 7.3-11
== cmdaemon ==
* Added: Docker Registry healthcheck
* Added: FreezeChangesToUGEConfig setting to the default cmd.conf file
* Improved: validation of Kubernetes role fields
* Fixed: an issue with Hadoop federation that can occur due to slowness of cm-hadoop-user script.
* Fixed: passing invalid members for chassis members crashes cmdaemon
* Fixed: removal of staticroute for bonded interface via cmsh neither removes the config file nor the route from memory
* Fixed: adding (slurm) jobqueues add queues with bad uniqueKeys
* Fixed: clone of a cloud node does not properly clone the EBS storage
* Fixed: export custom options to VSMP configuration
* Fixed: UGE parse qconf -sq pe_list forces comma separated list
* Fixed: rare crash in Event forwarding
* Fixed: updating CMDaemon on a passive head node makes the service fail to start
* Fixed: /etc/genders generated with incorrect expressions
* Fixed: pythoncm : Entity::usedby() returns empty list
* Fixed: cm-component-certificate --complete times out when the cluster has large number of nodes
* Fixed: IEEL version retrieval
== node-installer ==
* Updated: node-installer time sync. Node-installer now tries, after first trying to sync time 3 times the regular way, to set the time and date even if the ntp-daemon on the head node or net-master is not settled down.
== admin-manual ==
http://support.brightcomputing.com/manuals/7.3/admin-manual.pdf
* Added: advancedconfigs: CipherList, AllowNullCipherNetwork
* Added: wlm-setup --categories
* Updated: Kubernetes
* Updated: LSF version 10 support (cluster-tools r9905 onward)
== installation-manual ==
http://support.brightcomputing.com/manuals/7.3/installation-manual.pdf
* Added: cuda 8 installation
* Added: cluster-sync
== machine-learning-manual ==
http://support.brightcomputing.com/manuals/7.3/machine-learning-manual.pdf
* Added: sample digits/caffe run
* Added: tensorflow run examples
== cluster-tools ==
* Changed: Hadoop KMS will now use HTTPS protocol when enabling SSL via security script
* Changed: wlm-setup now by default assigns client roles only to the default category (new option --categories changes this behavior)
* Added: support for Apache Cassandra 3.9
* Added: support for Spark 2.0.1
* Added: support for Cloudera 5.8.2
* Added: support for Alluxio 1.3.0
* Added: support for Flink 1.1.3
* Added: support for Kafka 0.10.1.0
* Added: support for Cloudera 5.9.0
* Added: support for Hbase 0.98.23
* Added: LSF 10 support in wlm-setup
* Added: IEEL support for CentOS 7.2
* Improved: cm-kubernetes-setup reuse of existing and clean up of unused Kubernetes networks
* Updated: cm-docker-registry-setup to deploy registry healthchecks against the selected nodes.
* Updated: exclude list for IEEL
* Fixed: enable NetworkManager in lustre server image
* Fixed: an issue with conflicting packages in the 'cm' repo when setting up IEEL
== cm-apache-hadoop ==
* Added: support for Hbase 0.98.23
== cm-apache-hadoop-extras ==
* Added: support for Spark 2.0.1
* Added: support for Apache Cassandra 3.9
* Added: support for Kafka 0.10.1.0
== cm-cloudera-hadoop ==
* Added: support for Cloudera 5.8.2
* Added: support for Cloudera 5.9.0
== cmburn ==
* Fixed: an issue with kmon burn test failing on RHEL6
== cmgui ==
* Added: support for Spark 2.0.1
* Added: support for Apache Cassandra 3.9
* Added: support for Cloudera 5.8.2
* Added: support for Cloudera 5.7.4
* Added: support for Hbase 0.98.23
* Added: support for Cloudera 5.9.0
* Improved: Big Data deployment wizards user experience
* Improved: Kubernetes deployment wizard user experience
* Fixed: in some cases, an issue with duplicate serials preventing cmgui from connecting
* Fixed: request provisioning nodes update
* Fixed: uplinks ports for ethernet switch
== cuda-driver ==
* Updated: nvidia kernel driver to version 361.93.03
== intel-wired-ethernet-drivers ==
* Updated: igb driver to 5.3.5.4, i40e driver to 1.5.25, fm10k driver to 0.21.5.
== openstack-magnum ==
* Fixed: cm-docker-setup fails if openstack-magnum is installed in the software image
== pbspro ==
* Changed: install cgroup hooks by default (on new installations only)
== torque ==
* Upgrade to 6.0.2
== torque-slave ==
* Upgrade to 6.0.2