LSF Version 7.3 - Release Notes for Platform LSF

is programmed to report values more frequently than every 5 seconds, set the
ELIM_POLL_INTERVAL so that it samples information at a corresponding rate.
•
LSF_ELIM_BLOCKTIME=seconds—in Parameters section. UNIX only; used when the
external load indices feature is enabled. Maximum amount of time the master external
load information manager (MELIM) waits for a complete load update string from an
elim executable. After the time period specified by LSF_ELIM_BLOCKTIME, the MELIM
writes the last string sent by an elim in the LIM log file (lim.log.host_name) and
restarts the elim. Defining LSF_ELIM_BLOCKTIME also triggers the MELIM to restart
elim executables if the elim does not write a complete load update string within the time
specified for LSF_ELIM_BLOCKTIME.
•
LSF_ELIM_DEBUG=y—in Parameters section. UNIX only; used when the external load
indices feature is enabled. When this parameter is set to y, all external load information
received by the load information manager (LIM) from the master external load
information manager (MELIM) is logged in the LIM log file (lim.log.host_name).
Defining LSF_ELIM_DEBUG also triggers the MELIM to restart elim executables if the
elim does not write a complete load update string within the time specified for
LSF_ELIM_BLOCKTIME.
•
LSF_ELIM_RESTARTS=integer—in Parameters section. UNIX only; used when the
external load indices feature is enabled. Maximum number of times the master external
load information manager (MELIM) can restart elim executables on a host. Defining this
parameter prevents an ongoing restart loop in the case of a faulty elim. The MELIM waits
the LSF_ELIM_BLOCKTIME to receive a complete load update string before restarting
the elim. The MELIM does not restart any elim executables that exit with
ELIM_ABORT_VALUE.
Important:
Either LSF_ELIM_BLOCKTIME or LSF_ELIM_DEBUG must
also be defined; defining these parameters triggers the MELIM
to restart elim executables.
lsf.conf
•
LSF_PAM_CLEAN_JOB_DELAY=time_seconds—The number of seconds LSF waits
before killing a parallel job with failed tasks. Specifying LSF_PAM_CLEAN_JOB_DELAY
implies that if any parallel tasks fail, the entire job should exit without running the other
tasks in the job. The job is killed if any task exits with a non-zero exit code.Specify a value
greater than or equal to zero (0). Applies only to PAM jobs.
•
LSB_DEBUG_CMD adds LC_ADVRSV class to log advance reservation modifications
with brsvmod.
•
EGO_PREDEFINED_RESOURCES—When Platform EGO is enabled in the LSF cluster
(LSF_ENABLE_EGO=Y), you also can set the several EGO parameters related to LIM,
PIM, and ELIM in either lsf.conf or ego.conf. All clusters must have the same value
of EGO_PREDEFINED_RESOURCES in lsf.conf to enable the nprocs, ncores, and
nthreads host resources in remote clusters to be usable.
lsf.shared
A resource name cannot be any of the following reserved keywords:
cpu cpuf io logins ls idle maxmem maxswp maxtmp type model
status it mem ncpus nprocs ncores nthreads
define_ncpus_cores define_ncpus_procs define_ncpus_threads
ndisks pg r15m r15s r1m swap swp tmp ut
Release Notes for Platform LSF
Release Notes for Platform LSF 15