Installation guide

Release Notes
/etc/security/limits.conf
/etc/sysconfig/syslog
Additionally, Scyld ClusterWare uses /sbin/chkconfig to enable nfs.
Running nscd service on master node may cause kickbackdaemon to misbehave
The nscd (Name Service Cache Daemon) service executes by default on each compute node. However, if this service is also
enabled on the master node, then it may cause the Scyld ClusterWare name service kickbackdaemon to misbehave.
Workaround: when Beowulf starts, if it detects that nscd is running on the master node, then Beowulf automatically stops
nscd and reports that it has done so. Beowulf does not invoke /sbin/chkconfig nscd off to permanently turn off the service.
Note: even after stopping nscd on the master node,
/sbin/service nscd status
will report that nscd is running because the daemon continues to execute on each compute node, as controlled by
/etc/beowulf/init.d/09nscd.
Scyld ClusterWare MVAPICH CPU affinity management
CW4.2.0 (and later releases) support Infiniband via Open Source kernel drivers, OpenIB, OFED, and a Scyld ClusterWare-
enhanced MVAPICH. The CW4.2.0 MVAPICH default behavior is to assign threads of each multithreaded job to specific
CPUs in each node, starting with cpu0 and incrementing upward. While keeping threads pinned to a specific CPU may be
an optimal NUMA and CPU cache strategy for nodes that are dedicated solely to a single job, it is usually suboptimal if
multiple multithreaded jobs share a node, as each job’s threads get permanently assigned to the same low-numbered CPUs.
The CW4.2.1 (and beyond) default behavior is to not impose strict CPU affinity assignments, which allows the kernel CPU
scheduler to migrate threads as it sees fit to load-balance the node’s CPUs as workloads change over time.
However, the user may override this default using:
export VIADEV_ENABLE_AFFINITY=1
Conflicts with base distribution of openmpi and lam
Scyld ClusterWare 4.9.0 includes MPI-related packages that conflict with certain packages in the Red Hat or CentOS base
distribution.
If yum informs you that it cannot install or update Scyld ClusterWare because various mpich and mpiexec packages conflict
with various openmpi and lam packages from the base distribution, then run the command:
yum remove openmpi* lam*
to remove the conflicting base distribution packages, then retry the groupupdate of Scyld-ClusterWare.
Reducing the size of /usr/lib/locale/locale-archive
Glibc applications silently open the file /usr/lib/locale/locale-archive, which means it gets pulled into a node’s
libcache early in a node’s startup sequence. The default locale-archive is commonly many dozens of megabytes in size.
This consumes significant network bandwidth to move it from the master to each node, and thereafter consumes significant
23