Base Command Manager / Bright Cluster Manager Release Notes
Release notes for Bright 6.0-11 (released 2012-08-10)
buildmaster : /var/log/* is not excluded anymore when grabing a new image from a node on sles11.
buildmaster : The file /var/log/node-installer is now excluded when grabing a new image from node.
buildmaster : Added /etc/openvpn to default exclude list for update. Not having this would break cloud setups when doing imageupdate.
buildmaster : add missing dependencies (headnode: libidn-devel, image: net-snmp-perl, net-snmp) for RHEL5
cluster-tools : cmha-setup:improve failover config updates to store only relevant config
cluster-tools : cmhs-setup:do not ask for shared external iface params for type2 network topologies
cluster-tools : cmha-setup: format failover config file using xmllint if available
cluster-tools : align, make wlm-setup man page clear
cmdaemon : Improved text prompts in the request-licence utility.
cmdaemon : Added VLAN support on SLES. Note that setting the REORDER_HDR VLAN flag via the VLAN's ifcfg file is not supported by SLES.
cmdaemon : Added ability to specify full spool paths in cmd.conf: TorqueSpool, PBSProSpool.
cmdaemon : fix in interfaces healthcheck to ignore IB related warnings
cmdaemon : Improved the 'IP x.x.x.x is not contained in the network' validation error message to now contain the name of the interface
cmdaemon : pam authentication for json
cmdaemon : Fixed inaccurate help property descriptions for a service object's 'autostart' and 'monitored' properties.
cmdaemon : Fixed a small inconsistency in the cmsh's help screen for the 'show user' command.
cmdaemon : If a failing service is monitored, its status will now be automatically changed to 'UP' after the cmdaemon detects that the service is running again.
cmdaemon : Prior to this release the CMDaemon used to start services marked with monitoring=1 autostart=0 during the CMDaemon startup.
With this fix this is not the case anymore -- the service needs to be configured with autostart=1 monitoring=1 to be started
by the CMDaemon when the CMDaemon is starting.
cmdaemon : fix power status overview crash
cmdaemon : implemented option to log long soap calls
manuals : updated: net udev rules in base archive of cm-create-image
manuals : fixed: showport -s option in cmsh
manuals : improved: documentation for pxe install, cluster wide default image
manuals : improved: documentation clarification on metrics units and measurement units for cumulative metrics
manuals : updated: wlm-setup options
node-installer: Fixed 'SetupBmc' directive of the node-installer.conf having no effect (since 6.0 it used to incorrectly work as if it was always set to 'true',
regardless of the actual setting).
node-installer: Added VLAN support on SLES (provisioning over VLAN interfaces on SLES).
node-installer: Added support for installing a boot record on nodes that use software raid and/or LVM. This is for RHEL/CENTOS/SL-5/6 only.
node-installer: Fixed an issue where mkinitrd_cm would fail on RHEL HPC edition. Also, some non-server RHEL 6 variants were being detected as RHEL 5 variants by
the node installer, potentially leading to bad configuration file formats being written.
slurm : Changed the default value of SlurmctldTimeout from 5 to 0 in slurm.conf