Base Command Manager / Bright Cluster Manager Release Notes

Release notes for Bright 5.2-38 (released 2012-11-02)

== cmdaemon ==

* Added workload manager python calls
* Fixed false positive in mounts health check for lustre file systems
* Fixed IB related warnings in interfaces healthcheck
* Do not restart PBS Pro Mom service on cmdaemon restart
* Fixed cmd -t crash; cmd -t now prints to stdout
* Fixed cmha dbreclone handling of slurm configuration
* Fixed CMD_* not being set in metric scripts
* Fixed telnet shell taking 100% cpu
* Improved restart required messages

== node-installer ==

* Added /cm/node-installer/scripts/debug convenience script. The script can be used to toggle the debug mode of the node-installer.
* Fixed NFS not being mounted from the node-installer on diskless redhat 6 nodes

== node-installer-slave ==

* Use initramfs basename for RHEL6 (previously was initrd)
* Do not rely on build-config from primary during cm-clone-install

== manuals ==

* Improved cuda4.2 note
* Updated net udev rules in base archive of cm-create-image

== cciss_vol_status ==

* Update from 1.09 to 1.10, removed 1.09 patch

== finalise-base ==

* Added 0.9 metadata type to auto-assemble line of mdadm.conf, if not already configured.
* Resolves conflicts of multiple arch lm_sensors packages

== intel-compilers ==

* Update of Intel composer suite from 2011.11 to 2011.12

== intel-mpi ==

* Added I_MPI_ROOT enviroment variable to module files.

== intel-wired-ethernet-drivers ==

* Update of e1000e driver from 2.0.0.1 to 2.1.4, and ixgbe from 3.10.16 to 3.10.17
* Update of driver versions, igb: from 3.4.8 to 4.0.17, ixgbe: from 3.10.17 to 3.11.33, ixgbevf: from 2.6.2 to 2.7.12

== mlnx-ofed ==

* Added topdir config for building kernel RPMs
* Added a script to fix mlx4_en modprobe configuration if mlnx-ofed is already installed
* Update from 1.5.3-3.0.0 to 1.5.3-3.1.0, changed base archive used, from iso to tgz
* Renamed qlgc-ofed module name to mlnx-ofed

== slurm ==

* Changed the default value of SlurmctldTimeout from 5 to 0 in slurm.conf
* Patch applied: a ping of nodes which are in powersave mode fails