How to install AliEn Site

This document describes how to install and configure a head node on a site to support AliEn services. This is a node on which will be deployed long-lived agents and services and it is expected to be provided at the sites. The agents/services deployment and support is under the VO responsibility. The general requirements for this node and agents/services description is given below. Please visit Installation Guide topic to learn more details on how to download precompiled AliEn packages.

 

 

 

1 Prerequisites

 

  • General requirements
    • Linux kernel 2.6 or higher, any 64-bit Linux, preffered are SL6, Ubuntu
    • hardware: min. 4 GB RAM, 100GB of local (non mounted through shared FS) space.This space will be used for logs and services lock files.
  • Network connectivity

Port # Access: Outgoing + Service
==========================================================================

8084/TCP incoming from CERN and the site WN - ClusterMonitor

1093/TCP incoming from World - MonALISA FTD server, SE tests

8884/UDP incoming from the site WN and site SE nodes - Monitoring info

(only if not using CVMFS) 9991/TCP incoming from CERN - PackMan

9930/UDP incoming from the site SE nodes - Xrootd metrics

+   ICMP incoming and outgoing - network topology for file placement and access

==========================================================================

In general, the assumption is that the outgoing connectivity from the VO-box and the WNs is unrestricted.

 

  • Submission/query access to the site batrch system. AliEn supports PBS/Torque, LSF, SLURM, SGE.
  • Shared software area, in case the site does not support/allow software distribution through torrent

 

 

2 Software prerequisites

It is expected that the worker nodes have a standard Linux installation, including gcc. In addition, the ALICE application software requires gfortran. 

3. Site details

  • AliEn services operations are managed through the VO-box's user account
  • The user account should be accessible through a username/password or ssh keys
  • The local batch queue for the ALICE jobs should allow for at least 3GB of RAM and 4GB of swap space per slot, in addition every job requires ~10GB of scratch disk space during execution
  • The queue WALL time limit should be no less than 24 hours

2.2 Selecting Installer Options

Installer requires the dialog tool to be installed and will present you with a simple curses-based user interface. Here is what you need to select to install VO box:

 

  • Select latest stable version (or latest development version if you like to be up to date with developments)
  • Select binary installation type
  • No autodetect & reuse (this will allow to script to copy and install all necessary external software in

$VO_ALICE_SW_DIR directory.)

  • Select 'site', and 'monitor' components
  • Select directory for installation (on LCG VO-boxes: $VO_ALICE_SW_DIR/alien). You will need about 400 MB of disk space

 

3 Configuration

 

3.1 Setting up environemnt

On LCG VO boxes, add to .bashrc:

 

export VO_ALICE_SW_DIR=/lcg/directory/for/alice/software
export ALIEN_INS=$VO_ALICE_SW_DIR/alien

If your local Unix id is different from one known to VO LDAP server, define ALIEN_USER variable in $HOME/.alien/Environment file.

 

3.2 Preparing certificates

Put your certificate in $HOME/.globus and do

 

# cd $HOME/.alien
# ln -sf $HOME/.globus globus

 

3.3 Creating proxy certificate

Check that you can create proxy certificate

 

# alien proxy-init

 

 

3.4 Pairing a Site and VO

Once AliEn is installed, the site has to be paired with VO. To register your site, please send a mail to alice-project-alien-developers@cern.ch specifying:

 

  • Name of the institute (e.g. CERN)
  • Domain name (e.g. cern.ch)
  • Name of the city (e.g Geneva)
  • Name and e-mail of the administrator (e.g Pablo.Saiz@cern.ch)
  • Name of the front-end machine where AliEn services are installed (e.g. host.domain.org)
  • Path where to keep the log files (e.g. /home/alienMaster/AliEn/log)
  • Path in the machine to keep temporary files (e.g /tmp/AliEn/tmp)
  • Type of batch system (e.g. LSF,PBS,..)
  • Number of simultaneous jobs that can be executed for this V.O. (20)
  • Type of mass storage systme (Castor, HPSS, disk,...)
  • Maximum size of this buffer we are allowed to use
  • MSS path in which the files will be saved (e.g. /castor/cern.ch/user/a/aliprod/AliEn-1.0/data)
  • Host certificate subject
  • Maximum number of simultaneous file transfers to and from the site

Optional information:

 

  • Job working directory on the worker node (if not set by the batch system)
  • If the command to submit the jobs is not in the path, full path to the command
  • Any special arguments required for the submission of the job

When we receive this information, we will include it in our LDAP server, and you will find your site in the list of sites.

 

3.5 Registering your certificate with VO

If you are not already registered as AliEn user, do (only once):

 

# alien register-cert -user

If installation was sucessful, this should result in alien prompt.

 

# alien

You can connect as different user or to different VO by using appropriate command line switches:

 

# alien -user -org

 

4 Starting the services

The VO box should run the following services and monitoring agents:

 

  • Storage Element Service (SES) - interface to local storage (via SRM or directly)
  • File Transfer Daemon (FTD) - scheduled file transfers agent (possibly using FTS implementation)
  • xrootd - application file access
  • Cluster Monitor (CM) - local queue monitoring
  • MonALISA - general monitoring agent
  • PackMan (PM) - application software distribution and management
  • Computing Element (CE) - interface to local scheduler

This is how they can be started by hand:

 

# alien StartMonitor
# alien StartSE
# alien StartFTD
# alien StartPackMan
# alien StartMonaLisa
# alien StartCE

 

5 Checking that the services work

One all your services are running, there are a couple of easy test that you can try to check if the installation works properly

# echo "Hello world" >/tmp/myfile
# alien -exec add myNewFile /tmp/file
# alien login -exec packman list
# echo "Executable=\"date\";Requirements=other.CE==\" \" ">/tmp/jdl
# alien -exec add myJDL /tmp/jdl
# alien login -exec submit myJDL

The first two lines check that the SE works. The next line checks that the PackMan is working properly. Finally, a JDL is inserted in the catalogue and a job is submitted. After a while, your CE is supposed to execute it. If it doesn't, there is something wrong with the setup