Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 7.1-14
== cmdaemon ==
* OpenStack: fixed an issue with power operations on Bright Managed nodes. It used to not be possible to power them on, after powering them off.
* OpenStack: added the ability to specify (force) on which hypervisor host a bright managed node should be created. cmsh->vnode001->virtualsettings->defaultHost
* Added: adv. config option PXEBootOverHTTP to PXE boot nodes via cmdaemon HTTP instead of tftp
* Added: adv. config option ExtraRsyncOptions[.type].normal and ExtraRsyncOptions[.type].new-file for custom provisioning rsync options
* Improved: validation check to prevent spaces in the names of categories
* Fixed: in some cases, /var/spool/cmd/my.cmd.url becomming empty during HA failover
* Fixed: dhcpd.conf for VirtualSMP nodes
* Fixed: PBS Pro jobs are not listed in neither CMSH nor CMGUI
* Fixed: native jobs submitted with slurm to MIC cards failing and draining the node
* Fixed: an issue with named configuration when two networks share the same IP range
* Fixed: beegfs not automatically excluded during image update/grab
* Fixed: previously the interfaces healthcheck did not use the correct speed setting for some alias interfaces. Now all alias interfaces will be inherited from the parent interface.
* Fixed: rogueprocess healthcheck failing for sge_shepherd processes
== node-installer-slave ==
* Fixed: Bright managed OpenStack instances get stuck loading ldlinux.c32.
== manuals ==
* Added: PXEBootOverHTTP
* Added: bringupduringinstall
* Updated: [hadoop] updates
* Updated: [hadoop] updates
* Updated: [Hadoop]
* Updated: mounting the root filesystem for a diskless node over NFS - writable subtrees for CentOS 7 added
== cluster-tools ==
* Added: support for HAM (Slurm + Hadoop)
* Added: cmhadoop-hbase-setup (allows deployment/removal of HBase after Hadoop has been deployed)
* Added: support for Spark 1.4.1
* Added: support for Apache Drill
* Added: support for Apache Flink
* Added: support for Cloudera 5.3.6
* Added: support for Cloudera 5.4.5
* Improved: stability of the OpenStack deployment process
* Fixed: an issue with granting access to multiple Hadoop instances at the same time
== cm-apache-hadoop-extras ==
* Added: support for Spark 1.4.1
== cm-cloud-copy ==
* Fixed: keeping permissions when downloading file from S3 with different destination name
== cm-cloudera-hadoop ==
* Added: support for Cloudera 5.4.5
== cm-openstack-caas ==
* CaaS horizon dashboard
== cmgui ==
* Added: support for Cloudera 5.3.6
* Added: support for Cloudera 5.4.5
* Improved: disabled/enabled status of metrics/healthchecks is now shown in monitoring overview
* Fixed: an issue with removing network interfaces with cmgui if a bond interface is defined
* Fixed: changing the color settings from graphs
* Fixed: category node-service status overview
== cuda-driver ==
* Updated: NVIDIA driver to version 352.39.
== cuda75 ==
* Added: CUDA 7.5 packages
== intel-wired-ethernet-drivers ==
* Updated: igb driver from 5.3.2 to 5.3.2.2, ixgbe driver from 4.1.1 to 4.1.2, i40e driver from 1.2.48 to 1.3.38, i40evf driver from 1.2.34 to 1.3.25.
* Updated: e1000e driver from 3.1.0.2 to 3.2.4.2, igb driver from 5.2.18 to 5.3.2, ixgbe driver from 4.0.3 to 4.1.1, i40e driver from 1.2.38 to 1.2.48, i40evf driver from 1.2.25 to 1.2.34.
== sge ==
* Fixed: the GPU setting for sgeclient role not creating a GPU complex value