https://github.com/root-project/root
Raw File
Tip revision: 4009dcb8b8c9faa2ff0d013639b4056f0bed77fa authored by Gerardo Ganis on 26 October 2011, 11:28:57 UTC
Tip revision: 4009dcb
proof.conf.sample
# PROOF config file. It has a very simple format:
#
# master <hostname> [image=<imagename>] [workdir=<workingdirectory>]
# worker <hostname> [perf=<perfindex>] [image=<imagename>] [port=<portnumber>]
#                   [workdir=<workingdirectory>]
# submaster <hostname> [image=<imagename>] [port=<portnumber>]
#                      [config=<conffilename>]
# condorworker <virtualmachinename> [image=<imagename>] [perf=<perfindex>]
#                                   [workdir=<workingdirectory>]
# user <username> on <hostname>
#
# For each node in the cluster that can run a master add a "master" line.
# For each node in the cluster that can run a worker add a "worker" line.
# A single node can run multiple workers (e.g. an SMP machine). In that case
# repeat the worker line N times, where N is the number of CPU's in the machine.
# Perfindex is a number specifying the relative speed of each CPU in the
# cluster. Say, the fastest CPU is 100, a CPU 30% slower has perf=70. The
# perfindex is used for load balancing. By default perf=100. The perfindex
# is mostly an indication since the parallel algorithm is adaptive and will
# adjust itself to the current load on each machine.
# Image is used to tell the system that system images are shared between
# nodes/CPU's in the cluster. For example if all nodes share a common user
# directory, like with AFS, set image=afs (<imagename> is arbitrary but should
# be unique for each different image). By default imagename is the hostname.
# Port is used to tell PROOF to connect to the workers on the specified
# port (in case proofd was started by the user instead of via inetd).
#
# For condorworkers the special image IGNORE can be used to exclude
# a worker from the set of active workers. The worker (Condor VM) will
# be suspended and resumed. This supports situations where there a more
# VMs then CPUs.
# 
# Users can be redirected to have their master server run on a pre-determined
# node. For that use the "user" line.
#
# For backward compatibility it is still possible to specify the methods
# used for worker authentication as a list following the 'worker <hostname>'
# (e.g. worker <hostname> uidgid usrpwd srp), this is however deprecated.
# Authentication directives are specified via $ROOTSYS/etc/system.rootauthrc
# and/or $HOME/.rootauthrc files. See $ROOTSYS/etc/system.rootauthrc for details.
#

#master pcna49a image=nfs

#worker pcna49a  perf=100 image=nfs
#worker pcna49b  perf=100 image=nfs
#worker pcna49c  perf=100 image=nfs
#worker pcna49d  perf=100 image=nfs
#worker pcna49e  perf=75 image=nfs

#user rdm on pcna49d


master localhost

worker localhost
worker localhost
back to top