diff --git a/README b/README index 4a0eebcff790d93479e016d1b597efba408a39fd..93656788884315faaaa4c54acedaf1c814886687 100644 --- a/README +++ b/README @@ -58,71 +58,10 @@ quick description of the subdirectories of the SLURM distribution follows: COMPILING AND INSTALLING THE DISTRIBUTION ----------------------------------------- -Please the the INSTALL file for basic instructions. You will need a -working installation of OpenSSL. - -SLURM does not use reserved ports to authenticate communication -between components. You will need to have at least one "auth" -plugin. Currently, only three authentication plugins are available: -"auth/none," "auth/authd," and "auth/munge." The "auth/none" plugin is -built and used by default, but one of either Brent Chun's authd, or Chris -Dunlap's Munge should be installed in order to get properly authenticated -communications. The configure script in the top-level directory of this -distribution will determine which authentication plugins may be built. - - -OpenSSL: -http://www.openssl.org - -AUTHD: -http://www.theether.org/authd/ - -MUNGE: -http://www.llnl.gov/linux/munge/ - - -CONFIGURATION -------------- - -An annotated sample configuration file for SLURM is provided with this -distribution as etc/slurm.conf.example. Edit this config file to suit -your site and cluster, then copy it to `$sysconfdir/slurm.conf,' where -sysconfdir defaults to PREFIX/etc unless explicitly overwritten in the -`configure' or `make' steps. - -Once the config file is installed in the proper location, you'll need -to create the keys for SLURM job credential creation and verification. -The following openssl commands should be used: - - > openssl genrsa -out /path/to/private/key 1024 - > openssl rsa -in /path/to/private/key -pubout -out /path/to/public/key - -The private key and public key locations should be those specified by -JobCredentialPrivateKey and JobCredentialPublicCertificate in the SLURM -config file. - - -RUNNING SLURM -------------- - -Once a valid configuration has been set up and installed, the SLURM -controller, slurmctld, should be started on the primary and backup -control machines, and the SLURM compute node daemon, slurmd, should be -started on each compute server. - -The slurmd daemons need to run as root for production use, but may be -run as a user for testing purposes (obviously no jobs may be run as -any other user in that configuration). The SLURM controller, slurmctld, -need to be run as the configured SlurmUser (see your config file). - -Man pages are the best source of information about SLURM commands and -daemons. Please see: slurmctld(8), slurmd(8), scontrol(1), sinfo(1), -squeue(1), scancel(1), and srun(1). - -Also, take a look at the Quickstart Guide to get acquainted with -running and managing jobs with SLURM: doc/html/quickstart_admin.html -or PREFIX/share/doc/quickstart_admin.html. - +Please see the instructions at + http://www.llnl.gov/linux/slurm/quickstart_admin.html +Extensive documentation is available from our home page at + http://www.llnl.gov/linux/slurm PROBLEMS --------