Base Command Manager / Bright Cluster Manager Release Notes

Release notes for Bright 7.0-23

== cmdaemon ==

* Support for RHEL 7 and CentOS 7
* Changed: --hard-links as default option for rsync, can be turned off via AdvancedConfig RsyncHardLinks=0
* Changed: Turn off SSLv3 by default. To enable again AdvancedConfig: SSLv3=1
* Added: option for adding extra rsync options for image sync via exclude lists. for example: # extra-rsync-options: --max-delete=0
* Added: AdvancedConfig flags ExtraRsyncOptions.sync_mode for setting extra options for the rsync process
* Added: UCS force live info refresh
* Added: AdvancedConfig option to exclude hardware overview parameters, default=hostname,mac,tag. Add cpuspeed to fix dynamic cpu speed
* Added: AdvancedConfig option CustomUserCountScript to override the way cluster overview totalUsers and workingUsers is retrieved
* Added: AdvancedConfig: RsyncAttrs=-1(default) / 0(never) / 1(always) to use --xattrs option for rsync. Default is 1 on RHEL7 and SLES12, 0 otherwise
* Added: AdvancedConfig: RsyncAcls=-1(default) / 0(never) / 1(always) to use --xacls option for rsync. Default is 1 on RHEL7 and SLES12, 0 otherwise
* Improved: ignore /media when adding disk space metrics to the monitoring configuration
* Improved: InfiniBand (ib) healthcheck: ignore State and Physical state if the link layer is ethernet
* Improved: service autostart and chkconfig
* Improved: Clear restartRequired flag for newly added devices
* Fixed: pythoncm: device.checknow() not returning data if checks are run in certain order
* Fixed: possible head node cmdaemon crash in monitoring after cloning a node
* Fixed: possible cmdaemon crash when (disk space) metrics have parameters with blanks/space
* Fixed: heatmap: values equal to max could overflow and give intensity on the minimum

== node-installer ==

* Support for RHEL 7 and CentOS 7

== node-installer-slave ==

* Read clone config when performing simple clone

== manuals ==

* Added: OpenStackVXLANGroup directive
* Added: accumulo project
* Added: UGE installation and wlm-setup instructions
* Updated: dropped obsolete intel compiler suite v2011 support
* Updated: packaged supported Hadoop implementations for no-brainer installation

== buildmaster ==

* Added: multiple versions OFED stacks. As of this change, only the selected OFED stack will be installed

== cluster-tools ==

* cm-openstack-setup: openstack deployment config file is now saved in the log file
* cm-openstack-setup: better handling of terminal resizing
* cm-openstack-setup: configure passwordless ssh login for the nova user (needed for non-live instance migration and instance resizing)
* Hadoop: add support for Pivotal HD
* Added: LSF installation and integration support to wlm-setup
* Added: UGE installation and integration support to wlm-setup

== cm-apache-hadoop ==

* Update HBase from 0.98.6 to 0.98.6.1
* Updated: HBase 0.98.6.1 to 0.98.7

== cm-apache-hadoop-extras ==

* Updated: Accumulo 1.5.1 to 1.5.2 (for Hadoop 1) and Accumulo 1.6.0 to 1.6.1 (for Hadoop 2)

== cm-cloudera-hadoop ==

* Updated: CDH 5.1.3 to 5.2.0

== cm-config-intelcompliance-master ==

* Update because related packages are updated. intel-cluster-runtime is updated to version 3.7 and intel-cluster-checker is updated to version 2.2.1.

== cm-config-intelcompliance-slave ==

* Update because related packages are updated. intel-cluster-runtime is updated to version 3.7 and intel-cluster-checker is updated to version 2.2.1.

== cm-config-syslog ==

* The rsyslog on RHEL-based distros now uses local4 facility for all OpenStack logs

== cm-config-syslog-slave ==

* The rsyslog on RHEL-based distros now uses local4 facility for all OpenStack logs

== cm-hortonworks-hadoop ==

* HortonWorks HDP tarballs: HDP 1.3.8 and HDP 2.1.5

== cmgui ==

* Hadoop: add support for Pivotal HD
* Move temporary files from .out to .txt because Firefox 33 doesn't like the extension
* Added: pie chart
* Fixed: slots for all SLURM client categories being modified instead of the selected one
* Fixed: tasks which require confirmation could be performed on the wrong nodes
* Fixed: add/edit of new thresholds

== cuda-driver ==

* Update driver from version 340.29 to 340.32.

== intel-cluster-checker ==

* Update from version 2.1.2 to 2.2.1.

== intel-cluster-runtime ==

* Update from version 3.6 to 3.7.

== intel-mic ==

* Update MPSS from 3.3 to 3.4

== intel-tbb ==

* Update from version 42_20140601oss to 43_20140724oss.

== libpam_wlm ==

* Fixed: whitelist

== mlnx-ofed2 ==

* Updated: module paths for MPI packages in modules package. Modules which are used for MPI packages which are installed with the OFED stack.

== mlnx-ofed21 ==

* Updated: module paths for MPI packages in modules package. Modules which are used for MPI packages which are installed with the OFED stack.

== mlnx-ofed22 ==

* Updated: module paths for MPI packages in modules package. Modules which are used for MPI packages which are installed with the OFED stack.

== mlnx-ofed23 ==

* Updated: module paths for MPI packages in modules package. Modules which are used for MPI packages which are installed with the OFED stack.

== munge ==

* Fixed: munge.key corupption on Xeon Phi

== node-installer-nfsroot ==

* Updated: bash

== qlgc-ofed ==

* Updated: module paths for MPI packages in modules package. Modules which are used for MPI packages which are installed with the OFED stack.