Special Training Programme on Installing and Using High Performance Clusters

Administration

Material

wiki

Special Training Programme on Installing and Using High Performance Clusters | (smr 2121)

Go to day

Wednesday, 3 June 2009

09:30 - 10:00 Introduction to the training
- 09:30 Introduction to the training 30'
  
  Speaker: Clement Onime & Stefano Cozzini

10:00 - 10:45 linux advanced concepts and commands

-filesystem structure and permissions 
-usage of environment variables
-configuration of bash shell (.profile .bash_profile etc..)
-kernel space and user space distinction
-libraries; path search and dependency: associated environment variables 
-useful command tools: sort / grep / uniq and other. 

references:
see for instance:
http://tldp.org/LDP/GNU-Linux-Tools-Summary/html/GNU-Linux-Tools-Summary.html

10:00 linux advanced concepts and commands 45'

Speaker: Stefano Cozzini & Mangesh Patil (Democritos/Sissa eLab)

10:45 - 11:30 Lab: Exercise on linux command line

Exercises on remote machine where all the commands and tricks discussed previously can be played.

10:45 Lab: Exercise on linux command line 45'

12:00 - 12:45 Administering a Linux Workstation

item to be discussed:

- documentation: where to find help 
- linux distributions: RH based vs. debian based 
- tools for software installation: yum/rpm etc.. 
- network basic and configuration
- disk partitions and filesystems 
- review of /etc/init.d and chkconfig tool
- bootloaders: grub/lilo configuration 
- basic security issues

Location:

ICTP-INFN Laboratory (Mlab)

12:00 Administering a Linux Workstation 45'

Speaker: Antonio Messina (ICTP)

12:45 - 14:15 Lunch break
- 12:45 Lunch break 1h30'

14:15 - 16:00 Lab: install a Linux workstation and setup useful services

two installation methods:

-via CD: standard one and interactive.

-via NFS:
 see: http://www.democritos.it/hpc-wiki/index.php/School-2007/InstallingAndSettingUpYourOwnLinuxWorkstation 

server will in this case tocai.elab.sissa.it
kickstart will be loaded on tocai by users at the end of the first method.

Location:

ICTP-INFN Laboratory (Mlab)

14:15 Lab: install a Linux workstation and setup useful services 1h45'

16:00 - 16:30 break
- 16:00 break 30'
16:30 - 18:00 Tutorial on bash scripting
```
reference here:


http://www.democritos.it/hpc-wiki/index.php/Main/CrashCourseOnBashScripting
```
- 16:30 Tutorial on bash scripting 1h30'
  
  Speaker: eLab team as tutors
18:00 - 18:20 home work assignment
```
Each participant will be assigned with a small home-work to be done. 
```
Location: ICTP-INFN Laboratory (Mlab)
- 18:00 home work assignment 20'

Thursday, 4 June 2009
- 09:00 - 09:45 Introduction to HPC concepts
  - 09:00 Introduction to HPC concepts 45'
    
    Speaker: Stefano Cozzini (Democritos/Sissa eLab)
- 09:45 - 10:30 HPC tools for Linux workstation
  - 09:45 HPC tools for Linux workstation 45'
    
    Speaker: Stefano Cozzini
- 10:30 - 11:00 break
  - 10:30 break 30'
- 11:00 - 12:30 Lab: installing tools for HPC on your workstation
```
Compilers, modules and libraries

references:
for compilers:

http://www.democritos.it/hpc-wiki/index.php/Main/InstallingTheGNUPGIAndIntelCompilerPackages

for libraries: ( to be checked)
http://www.democritos.it/hpc-wiki/index.php/Main/ComparePerformancesOfDifferentOptimizedLibrariesOnMatrixMultiplication

-We need to add how to install modules and how to create modules 
```
  - 11:00 Lab: installing tools for HPC on your workstation 1h30'
- 12:30 - 14:00 Lunch break
  - 12:30 Lunch break 1h30'
- 14:00 - 14:45 Advanced topic on sys. adm. for HPC
```
-advance storage configuration:
      -lvm and raid (hardware/software)

-troubleshooting 

-advance ssh usage
```
  - 14:00 Advanced topic on sys. adm. for HPC 45'
    
    Speaker: Piero Calucci & Antonio Messina
- 14:45 - 16:00 Lab: configure an I/O server for HPC
```
see here for reference:


http://www.democritos.it/hpc-wiki/index.php/School-2007/InstallAndBenchmarkASmallIOServer
```
  - 14:45 Lab: configure an I/O server for HPC 1h15'
- 16:00 - 16:30 break
  - 16:00 break 30'
- 16:30 - 18:00 Participants' contributions
```
Each participant will present himself and show his homework
(10 minutes each) 
```
  - 16:30 Participants' contributions 1h30'
Friday, 5 June 2009
- 09:00 - 09:45 Cluster Computing Concepts
  - 09:00 Cluster Computing Concepts 45'
    
    Speaker: Stefano Cozzini
- 09:45 - 10:30 the Npaci-Rocks approach to cluster computing
  - 09:45 the Npaci-Rocks approach to cluster computing 45'
    
    Speaker: Mangesh Patil
- 10:30 - 11:00 break
  - 10:30 break 30'
- 11:00 - 12:30 Lab: installing an Npaci-Rock cluster
  - 11:00 Lab: installing an Npaci-Rock cluster 1h30'
    
    Speaker: tutor: Mangesh and Alwaaled
- 12:30 - 14:00 Lunch break
  - 12:30 Lunch break 1h30'
- 14:00 - 14:30 Managing a Npaci-Rocks cluster
```
-troubleshooting 
-configure a second network interface
-add/remove software 
```
  - 14:00 Managing a Npaci-Rocks cluster 30'
    
    Speaker: Mangesh Patil
- 14:30 - 16:00 Lab: play with Npaci-rocks cluster
```
- install compilers 
- install modules
- manage software with module 
- configure the second network
- benchmark your cluster 
- (advanced: install myrinet) 
```
  - 14:30 Lab: play with Npaci-rocks cluster 1h30'
- 16:00 - 16:30 break
  - 16:00 break 30'
- 16:30 - 17:15 Wrap-Up session
```
Qustions/Answers Session on the program so far.
Homework assignement for the week-end
```
  - 16:30 Wrap-Up session 45'
Monday, 8 June 2009
- 09:00 - 09:45 Modern HPC platform for Linux Clusters
  - 09:00 Modern HPC platform for Linux Clusters 45'
    
    Speaker: Stefano Cozzini
- 09:45 - 10:30 Advanced Linux Cluster installation procedure (I)
  - 09:45 Advanced Linux Cluster installation procedure (I) 45'
    
    Speaker: Moreno Baricevic
- 10:30 - 11:00 Break
  - 10:30 Break 30'
- 11:00 - 11:30 Advanced Linux cluster installation procedure(II)
  - 11:00 Advanced Linux cluster installation procedure(II) 30'
    
    Speaker: Moreno Baricevic
- 11:30 - 13:00 Lab: installing a cluster (part 1)
```
reference: 

http://www.democritos.it/hpc-wiki/index.php/Main/ClusterInstallationPart1
```
  - 11:30 Lab: installing a cluster (part 1) 1h30'
- 13:00 - 14:30 Lunch Break
  - 13:00 Lunch Break 1h30'
- 14:30 - 16:00 Lab: cluster installation (part II)
  - 14:30 Lab: cluster installation (part II) 1h30'
- 16:00 - 16:30 Break
  - 16:00 Break 30'
- 16:30 - 17:30 Lab: cluster installation (part 3)
  - 16:30 Lab: cluster installation (part 3) 1h0'
- 17:30 - 18:00 Wrap-up session
  - 17:30 Wrap-up session 30'
Tuesday, 9 June 2009
- 09:00 - 09:45 Batch systems: introduction and configurations
  - 09:00 Batch systems: introduction and configurations 45'
    
    Speaker: Piero Calucci
- 09:45 - 10:05 Cluster Management: part 3
  - 09:45 Cluster Management: part 3 20'
    
    Speaker: Moreno Baricevic
- 10:05 - 10:30 Nagios tool for cluster monitoring
  - 10:05 Nagios tool for cluster monitoring 25'
- 10:30 - 11:00 break
  - 10:30 break 30'
- 11:00 - 12:30 Lab: Cluster installation part 2
```
reference:

http://www.democritos.it/hpc-wiki/index.php/Main/ClusterInstallationPart2
```
  - 11:00 Lab: Cluster installation part 2 1h30'
- 12:30 - 14:00 Lunch break
  - 12:30 Lunch break 1h30'
- 14:00 - 16:00 Lab: Cluster installation part 2
  - 14:00 Lab: Cluster installation part 2 2h0'
- 16:00 - 16:30 break
  - 16:00 break 30'
- 16:30 - 18:00 Wrap-up session
```
Questions/Answers/Comments etc. etc.
```
  - 16:30 Wrap-up session 1h30'
Wednesday, 10 June 2009
- 09:00 - 09:45 Parallel Filesystem: Lustre
  - 09:00 Parallel Filesystem: Lustre 45'
    
    Speaker: Piero Calucci
- 09:45 - 10:30 Advanced cluster networking
```
-channel bonding 
-myrinet 
-infiniband
-how to measure performances
-troubleshooting  
```
  - 09:45 Advanced cluster networking 45'
    
    Speaker: to be assigned
- 10:30 - 11:00 break
  - 10:30 break 30'
- 11:00 - 13:00 Lab: channel bonding / Vlan etc.
```
http://www.democritos.it/hpc-wiki/index.php/Main/ClusterInstallationPart4
```
  - 11:00 Lab: channel bonding / Vlan etc. 2h0'
- 13:00 - 14:00 Lunch Break
  - 13:00 Lunch Break 1h0'
- 14:00 - 16:00 Lab: Install and benchmark a small Myrinet network.
```
references:


http://www.democritos.it/hpc-wiki/index.php/School-2007/HighSpeedNetworkInstallationAndBenchmarkingMyrinet
```
  - 14:00 Lab: Install and benchmark a small Myrinet network. 2h0'
- 16:00 - 16:30 Break
  - 16:00 Break 30'
- 16:30 - 18:30 Lab: move NFS over myrinet and benchmark I/O performance
```
To be done 
```
  - 16:30 Lab: move NFS over myrinet and benchmark I/O performance 2h0'