Installation guide

Release Notes
RAM filesystem space on the node. It is likely that your cluster’s users and applications do not require all the international
locale data that is present in the default file. With care, the cluster administrator may choose to rebuild locale-archive
with a greatly reduced set of locales. For example, on a quiescent master node:
cd /usr/lib
mv locale locale.orig
mkdir locale
cp -a locale.orig/en_US* locale
/usr/sbin/build-locale-archive
rebuilds /usr/lib/locale/locale-archive as a 2MB file with only the locale data for United States English, and
preserves the original /usr/lib/locale directory as a backup.
Beofdisk does not support local disks without partition tables
Currently, beofdisk only supports disks that already have partition tables, even if those tables are empty. Compute nodes
with preconfigured hardware RAID, where partition tables have been created on the LUNs, should be configurable. Contact
Customer Service for assistance with a disk without partition tables.
Master node or compute node reports "instable time" in syslog
The Linux 2.6.9 kernel’s x86_64 timer code is overly aggressive in declaring "lost ticks" and for some workloads produces
a spurious warning, such as "Many lost ticks. Your time source seems instable or some driver is hogging interrupts." This
message can be safely ignored.
Issues with bproc and the getpid() syscall
BProc interaction with getpid() may return incorrect processID values.
Details: The Red Hat’s glibc implements the getpid() syscall by asking the kernel once for the current processID value, then
caching that value for subsequent calls to getpid(). If a program calls getpid() before calling bproc_rfork() or bproc_vrfork(),
then bproc silently changes the child’s processID, but a subsequent getpid() continues to return the former cached processID
value.
Workaround: do not call getpid() prior to calling bproc_[v]rfork.
Issues with ptrace
ptrace-based debuggers may fail to re-attach.
Details: You may use a ptrace-based debugger, such as gdb, running on a master node to debug a process that is running
on a compute node. However, if the debugger exits without cleanly detaching from the traced process using ptrace_detach,
then any further attempts to do a ptrace_attach to the target process will fail. For example, if gdb is tracing a process on a
compute node and gdb is killed by a signal, then further tracing of that target process will not be possible.
24