OAR/ installation/ Installation for the 2.5.x versions

Table of content:

  1. Installing the OAR batch system
    1. Overview
    2. Computing nodes
      1. Installation from the packages
      2. Installation from the tarball
      3. Configuration
    3. Server node
      1. Installation from the packages
      2. Installation from the tarball
      3. Configuration
      4. Notes
    4. Frontend nodes
      1. Installation from the packages
      2. Installation from the tarball
      3. Configuration
      4. OAR RESTful API Installation
    5. Visualization node
      1. Description
      2. Installation from the packages
      3. Installation from the tarball
      4. Configuration
    6. Further informations

Installing the OAR batch system

Overview

There are currently 3 methods to install OAR (All of them are documented in this page) :

  • with the debian packages
  • with the rpm packages
  • with the sources

The first thing you have to know is about the OAR architecture. A common OAR installation is composed of:

  • a server node which will hold all of OAR "smartness". This node will run the oar server daemon;
  • frontend nodes on which you will be allowed to login, then reserve some computing nodes (oarsub, oarstat, oarnodes, ...);
  • several computing nodes (a.k.a. the nodes), on which the jobs will run.
  • and optionally a visualisation node on which all the visualisation web interfaces (monika, draw-gantt, ...) will be accessible ;

Computing nodes

Installation from the packages

Instructions

For redhat like systems:

# Add the oar repository
cat <<EOF > /etc/yum.repos.d/oar.repo
[oar]
name=OAR Packages for Enterprise Linux 6 - \$basearch
baseurl=http://oar-ftp.imag.fr/oar/2.5/rpm/stable/
enabled=1
gpgcheck=0
EOF

# Install OAR node
yum install oar-node

For the debian like systems:

# Add the OAR repository (choose the right one. See http://oar.imag.fr/repositories/)
echo "deb http://oar-ftp.imag.fr/oar/2.5/debian squeeze main" > /etc/apt/sources.list.d/oar.list
curl http://oar-ftp.imag.fr/oar/oarmaster.asc | sudo apt-key add -
apt-get update

# Install OAR node
apt-get install oar-node

Installation from the tarball

Requirements

For redhat like systems:

# Build dependencies
yum install gcc make tar python-docutils

# Common dependencies
yum install Perl Perl-base openssh

For debian like system:

# Build dependencies
apt-get install gcc make tar python-docutils

# Common dependencies
apt-get install perl perl-base openssh-client openssh-server

Instructions

Get the sources:

OAR_VERSION=2.5.2
curl http://oar-ftp.imag.fr/oar/2.5/sources/stable/oar-${OAR_VERSION}.tgz | tar xzvf -
cd oar-${OAR_VERSION}/

build/install/setup:

# build
make node-build
# install
make node-install
# setup
make node-setup

Configuration

oar node ssh access

You need to ensure that the oar user can access to each nodes through ssh. To ensure that, you can just copy the /var/lib/oar/.ssh folder from the oar server to each nodes (ensure that /var/lib/oar/.ssh has the right permissions).

Init.d scripts

If you have installed OAR from sources, you need to become root user and install manually the {init.d,default,sysconfig} scripts present in the folders:

$PREFIX/share/doc/oar-node/examples/scripts/{init.d,default,sysconfig}

Then you just need to use the script /etc/init.d/oar-node to start the ssh daemon dedicated to oar-node.

Server node

Installation from the packages

Instructions

For redhat like systems:

# Add the epel repository (choose the right version depending on your operating system)
rpm -i http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-5.noarch.rpm

# Add the oar repository
cat <<EOF > /etc/yum.repos.d/oar.repo
[oar]
name=OAR Packages for Enterprise Linux 6 - \$basearch
baseurl=http://oar-ftp.imag.fr/oar/2.5/rpm/stable/
enabled=1
gpgcheck=0
EOF

# Install OAR server for the PostgreSQL backend
yum install oar-server oar-server-pgsql

# or Install OAR server for the MySQL backend
yum install oar-server oar-server-mysql

For the debian like systems:

# Add the OAR repository (choose the right one. See http://oar.imag.fr/repositories/)
echo "deb http://oar-ftp.imag.fr/oar/2.5/debian squeeze main" > /etc/apt/sources.list.d/oar.list
curl http://oar-ftp.imag.fr/oar/oarmaster.asc | sudo apt-key add -
apt-get update

# Install OAR server for the PostgreSQL backend
apt-get install oar-server oar-server-pgsql

# or Install OAR server for the MySQL backend
apt-get install oar-server oar-server-mysql

Installation from the tarball

Requirements

For redhat like systems:

# Add the epel repository (choose the right version depending on your operating system)
rpm -i http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-5.noarch.rpm

# Build dependencies
yum install gcc make tar python-docutils

# Common dependencies
yum install Perl Perl-base openssh Perl-DBI perl-Sort-Versions

# MySQL dependencies
yum install mysql-server mysql perl-DBD-MySQL

# PostgreSQL dependencies
yum install postgresql-server postgresql perl-DBD-Pg

For debian like system:

# Build dependencies
apt-get install gcc make tar python-docutils

# Common dependencies
apt-get install perl perl-base openssh-client openssh-server libdbi-perl libsort-versions-perl

# MySQL dependencies
apt-get install mysql-server mysql-client libdbd-mysql-perl

# PostgreSQL dependencies
apt-get install postgresql-server postgresql-client libdbd-pg-perl

Instructions

Get the sources:

OAR_VERSION=2.5.2
curl http://oar-ftp.imag.fr/oar/2.5/sources/stable/oar-${OAR_VERSION}.tgz | tar xzvf -
cd oar-${OAR_VERSION}/

Build/Install/Setup the OAR server:

# build
make server-build
# install
make server-install
# setup
make server-setup

Configuration

The oar database

Define the database configuration in /etc/oar/oar.conf. You need to set the variables DB_TYPE, DB_HOSTNAME, DB_PORT, DB_BASE_NAME, DB_BASE_LOGIN, DB_BASE_PASSWD, DB_BASE_LOGIN_RO, DB_BASE_PASSWD_RO:

vi /etc/oar/oar.conf

Create the database and the database users:

# General case
oar-database --create --db-admin-user <ADMIN_USER> --db-admin-pass <ADMIN_PASS>

# OR, for PostgreSQL, in case the database is installed locally
oar-database --create --db-is-local

Init.d scripts

If you have installed OAR from sources, you need to become root user and install manually the init.d/default/sysconfig scripts present in the folders:

$PREFIX/share/doc/oar-server/examples/scripts/{init.d,default,sysconfig}

Then use the script /etc/init.d/oar-server to start the OAR server daemon.

Adding resources to the system

If you want to automatically initialize your cluster then you just need to launch oar_resources_init. It will detect the resources from the nodes that you put in a file and store right OAR commands to initialize the database with the appropriate values for the memory and the cpuset properties. Just try...

There is also a tool to help you managing your oar resources and admission rules : oaradmin. Take a look at the oaradmin documentation in the administrator commands section for more details. You can also read this tips:

http://oar.imag.fr/archive/wiki-oar/index.php/Customization_tips#Using_oaradmin_to_initiate_the_resources

Otherwise:

To add resources to your system, you can use (as root) the command oarnodesetting. For a complete comprehension of what does this command, type man oarnodesetting. For now, the two options you will need will be -a (means add a resource) and -h (defines the resource hostname or ip adress).

For example, to add a computing resource on the node to OAR installation, you can type:

oarnodesetting -a -h <NODE_IP>

This will add a resource with as host IP address.

You also can modify resources properties with -p option, for example:

oarnodesetting -r 1 -p "deploy=YES"

will allow the resource #1 to accept jobs of the type deploy.

Notes

Security issues

For security reasons it is hardly recommended to configure a read only account for the OAR database (like the above example). Thus you will be able to add this data in DB_BASE_LOGIN_RO and DB_BASE_PASSWD_RO in oar.conf.

PostgreSQL : autovacuum

Be sure to activate the "autovacuum" feature in the "postgresql.conf" file (OAR creates and deletes a lot of records and this setting cleans the postgres database from unneeded records).

PostgreSQL : authentication

In case you've installed a PostgreSQL database remotly, if your PostgreSQL installation doesn't authorize the local connections by default, you need to enable the connections to this database for the oar users. Supposing the OAR server has the address , you can add the following lines in the pg_hba.conf:

# in /etc/postgresql/8.1/main/pg_hba.conf or /var/lib/pgsql/data/pg_hba.conf ~ host oar oar_ro /32 md5 host oar oar /32 md5

About X11 usage in OAR

The easiest and scalable way to use X11 application on cluster nodes is to open X11 ports and set the right DISPLAY environment variable by hand. Otherwise users can use X11 forwarding via ssh to access cluster frontal. After that you must configure ssh server on this frontal with

X11Forwarding yes
X11UseLocalhost no

With this configuration, users can launch X11 applications after a 'oarsub -I' on the given node or "oarsh -X node12".

Using Taktuk

If you want to use taktuk to manage remote administration commands, you have to install it. You can find information about taktuk from its website: http://taktuk.gforge.inria.fr.

Note: Taktuk is scalable remote command execution without the need to install special stuffs on nodes. So it is very useful to administer a large amount of server.

Then, you have to edit your oar configuration file and to fill in the different related parameters:

  • TAKTUK_CMD (the path to the taktuk command)
  • PINGCHECKER_TAKTUK_ARG_COMMAND (the command used to check resources states)
  • SCHEDULER_NODE_MANAGER_SLEEP_CMD (command used for halting nodes)

CPUSET feature

OAR uses the CPUSET features provided with the Linux kernel >= 2.6. This enables to restrict user processes only on reserved processors and to clean correctly the nodes after the end of the jobs.

For more information, look at the CPUSET file.

Energy saving

Starting with version 2.4.3, OAR provides a module responsible of advanced management of wake-up/shut-down of nodes when they are not used. To activate this feature, you have to:

  • provide 2 commands or scripts which will be executed on the oar server to shutdown (or set into standby) some nodes and to wake-up some nodes (configure the path of those commands into the ENERGY_SAVING_NODE_MANAGER_WAKE_UP_CMD and ENERGY_SAVING_NODE_MANAGER_SHUT_DOWN_CMD variables into oar.conf)

  • configure the "available_upto" property of all your nodes:

    • available_upto=0 : to disable the wake-up and halt
    • available_upto=1 : to disable the wake-up (but not the halt)
    • available_upto=2147483647 : to disable the halt (but not the wake-up)
    • available_upto=2147483646 : to enable wake-up/halt forever
    • available_upto= : to enable the halt, and the wake-up until the date given by

  • activate the energy saving module by setting ENERGY_SAVING_INTERNAL="yes" and configuring the ENERGY_* variables into oar.conf

  • configure the metascheduler time values into SCHEDULER_NODE_MANAGER_IDLE_TIME, SCHEDULER_NODE_MANAGER_SLEEP_TIME and SCHEDULER_NODE_MANAGER_WAKEUP_TIME variables of the oar.conf file.
  • restart the oar server (you should see an "Almighty" process more).

You need to restart OAR each time you change an ENERGY_* variable. More informations are available inside the oar.conf file itself. For more details about the mechanism, take a look at the "Hulot" module documentation.

Disabling SELinux

On some distributions, SELinux is enabled by default. There is currently no OAR support for SELinux. So, you need to disable SELinux, if enabled.

Intel cpuset id issue

The cpuset ids on an intel platform are not persistent across reboot. So you need to update the cpuset ids in the resource database at startup for each computing node. You can do this by using the /etc/oar/update_cpuset_id.sh script. The following page give more informations on how configuring it:

http://oar.imag.fr/archive/wiki-oar/index.php/Configuration_tips#Start.2Fstop_of_nodes_using_ssh_keys

Other issues

You can take a look at the "Customizaion tips" on the OAR Wiki:

http://oar.imag.fr/archive/wiki-oar/index.php/Customization_tips

Frontend nodes

Installation from the packages

Instructions

For redhat like systems:

# Add the epel repository (choose the right version depending on your operating system)
rpm -i http://download.fedoraproject.org/pub/epel/6/x86_64/epel-release-6-5.noarch.rpm

# Add the oar repository
cat <<EOF > /etc/yum.repos.d/oar.repo
[oar]
name=OAR Packages for Enterprise Linux 6 - \$basearch
baseurl=http://oar-ftp.imag.fr/oar/2.5/rpm/stable/
enabled=1
gpgcheck=0
EOF

# Install OAR user for the PostgreSQL backend
yum install oar-user oar-user-pgsql

# or Install OAR user for the MySQL backend
yum install oar-user oar-user-mysql

For the debian like systems:

# Add the OAR repository (choose the right one. See http://oar.imag.fr/repositories/)
echo "deb http://oar-ftp.imag.fr/oar/2.5/debian squeeze main" > /etc/apt/sources.list.d/oar.list
curl http://oar-ftp.imag.fr/oar/oarmaster.asc | sudo apt-key add -
apt-get update

# Install OAR server for the PostgreSQL backend
apt-get install oar-user oar-user-pgsql

# or Install OAR server for the MySQL backend
apt-get install oar-user oar-user-mysql

Installation from the tarball

Requirements

For redhat like systems:

# Build dependencies
yum install gcc make tar python-docutils

# Common dependencies
yum install Perl Perl-base openssh Perl-DBI

# MySQL dependencies
yum install mysql perl-DBD-MySQL

# PostgreSQL dependencies
yum install postgresql perl-DBD-Pg

For debian like system:

# Build dependencies
apt-get install gcc make tar python-docutils

# Common dependencies
apt-get install perl perl-base openssh-client openssh-server libdbi-perl

# MySQL dependencies
apt-get install mysql-client libdbd-mysql-perl

# PostgreSQL dependencies
apt-get install postgresql-client libdbd-pg-perl

Instructions

Get the sources:

OAR_VERSION=2.5.2
curl http://oar-ftp.imag.fr/oar/2.5/sources/stable/oar-${OAR_VERSION}.tgz | tar xzvf -
cd oar-${OAR_VERSION}/

Build/Install/setup:

# build
make user-build
# install
make user-install
# setup
make user-setup

Configuration

Coherent configuration files between server node and user nodes

You need to have a coherent oar configuration between the server node and the user nodes. So you can just copy the /etc/oar directory from to server node to the user nodes.

OAR RESTful API Installation

Since the version 2.5.2, OAR offers an API for users and admins interactions. This api must be installed on a frontend node (with the user module installed).

From the packaging

For redhat like systems:

# Add the oar repository
cat <<EOF > /etc/yum.repos.d/oar.repo
[oar]
name=OAR Packages for Enterprise Linux 6 - \$basearch
baseurl=http://oar-ftp.imag.fr/oar/2.5/rpm/stable/
enabled=1
gpgcheck=0
EOF

# Install apache FastCGI module (optional but highly recommended)
FIXME:

# Install OAR Restful api
yum install oar-restful-api

For the debian like systems:

# Add the OAR repository (choose the right one. See http://oar.imag.fr/repositories/)
echo "deb http://oar-ftp.imag.fr/oar/2.5/debian squeeze main" > /etc/apt/sources.list.d/oar.list
curl http://oar-ftp.imag.fr/oar/oarmaster.asc | sudo apt-key add -
apt-get update

# Install apache FastCGI module (optional but highly recommended)
apt-get install libapache2-mod-fastcgi 

# Install OAR Restful api
apt-get install oar-restful-api

From the sources

Requirements:

For redhat like systems:

# Build dependencies
yum install gcc make tar python-docutils

# Common dependencies
yum install perl perl-base perl-DBI perl-CGI perl-JSON perl-YAML perl-libwww-perl httpd

# FastCGI dependency (optional but highly recommended)
FIXME:

# MySQL dependencies
yum install mysql perl-DBD-MySQL

# PostgreSQL dependencies
yum install postgresql perl-DBD-Pg

For debian like system:

# Build dependencies
apt-get install gcc make tar python-docutils

# Common dependencies
apt-get install perl perl-base libdbi-perl libjson-perl libyaml-perl libwww-perl httpd-cgi libcgi-fast-perl 

# FastCGI dependency (optional but highly recommended)
apt-get install libapache2-mod-fastcgi

# MySQL dependencies
apt-get install mysql-server mysql-client libdbd-mysql-perl

# PostgreSQL dependencies
apt-get install postgresql-server postgresql-client libdbd-pg-perl

Instructions

Get the sources:

OAR_VERSION=2.5.2
curl http://oar-ftp.imag.fr/oar/2.5/sources/stable/oar-${OAR_VERSION}.tgz | tar xzvf -
cd oar-${OAR_VERSION}/

build/install/setup:

# build
make api-build
# install
make api-install
# setup
make api-setup

Configuration

Configuring OAR

For the moment, the API needs the user tools to be installed on the same host ('make user-install' or oar-user package). A suitable /etc/oar/oar.conf should be present. For the API to work, you should have the oarstat/oarnodes/oarsub commands to work (on the same host you installed the API)

Configuring Apache

The api provides a default configuration file (/etc/oar/apache-api.conf) that is using a identd user identification enabled only from localhost. Edit the /etc/oar/apache-api.conf file and customize it to reflect the authentication mechanism you want to use. For ident, you may have to install a "identd" daemon on your distrib. The steps may be:

  • Install and run an identd daemon on your server (like pidentd).
  • Activate the ident auth mechanism into apache (a2enmod ident).
  • Activate the headers apache module (a2enmod headers).
  • Activate the rewrite apache module (a2enmod rewrite).
  • Customize apache-api.conf to allow the hosts you trust for ident.

YAML, JSON, XML

You need at least one of the YAML or JSON perl module to be installed on the host running the API.

Test

You may test the API with a simple wget:

wget -O - http://localhost/oarapi/resources.html

It should give you the list of resources in the yaml format but enclosed in an html page. To test if the authentication works, you need to post a new job. See the example.txt file that gives you example queries with a ruby rest client.

Visualization node

Description

There are two different tools. One, named Monika which displays the current cluster state with all active and waiting jobs. The other, named drawgantt which displays node occupation in a lapse of time. These tools are CGI scripts and generate HTML pages.

Installation from the packages

Instructions

For redhat like systems:

# Add the oar repository
cat <<EOF > /etc/yum.repos.d/oar.repo
[oar]
name=OAR Packages for Enterprise Linux 6 - \$basearch
baseurl=http://oar-ftp.imag.fr/oar/2.5/rpm/stable/
enabled=1
gpgcheck=0
EOF

yum install oar-web-status

For the debian like systems:

# Add the OAR repository (choose the right one. See http://oar.imag.fr/repositories/)
echo "deb http://oar-ftp.imag.fr/oar/2.5/debian squeeze main" > /etc/apt/sources.list.d/oar.list
curl http://oar-ftp.imag.fr/oar/oarmaster.asc | sudo apt-key add -
apt-get update

apt-get install oar-web-status

Installation from the tarball

Requirements:

For redhat like systems:

# Build dependencies
yum install gcc make tar python-docutils

# Common dependencies
yum install perl perl-base perl-DBI ruby-GD ruby-DBI perl-Tie-IxHash perl-Sort-Naturally perl-AppConfig

# MySQL dependencies
yum install mysql perl-DBD-MySQL ruby-mysql

# PostgreSQL dependencies
yum install postgresql perl-DBD-Pg ruby-pg

For debian like system:

# Build dependencies
apt-get install gcc make tar python-docutils

# Common dependencies
apt-get install perl perl-base ruby libgd-ruby1.8 libdbi-perl libtie-ixhash-perl libappconfig-perl libsort-naturally-perl

# MySQL dependencies
apt-get install libdbd-mysql-perl libdbd-mysql-ruby

# PostgreSQL dependencies
apt-get install libdbd-pg-perl libdbd-pg-ruby

Instructions

Get the sources:

OAR_VERSION=2.5.2
curl http://oar-ftp.imag.fr/oar/2.5/sources/stable/oar-${OAR_VERSION}.tgz | tar xzvf -
cd oar-${OAR_VERSION}/

build/install/setup:

# build
make monika-build draw-gantt-build www-conf-build
# install
make monika-install draw-gantt-install www-conf-install
# setup
make monika-setup draw-gantt-setup www-conf-setup

Configuration

Drawgantt configuration

  • Edit /etc/oar/drawgantt.conf to fit your configuration.

Monika configuration

  • Edit /etc/oar/monika.conf to fit your configuration.

httpd configuration

  • You need to edit /etc/oar/apache.conf to fit your needs and verify that you http server configured.

Further informations

For further information, please check the documentation section on the OAR website http://oar.imag.fr/.