UbuntuHelp:MpichCluster

文章出处:

https://help.ubuntu.com/community/MpichCluster

点击翻译:

English

请不要直接编辑翻译本页，本页将定期与来源同步。

Setting Up an MPICH2 Cluster in Ubuntu

Creator: Omid Alemi [email protected] This guide describes how to building a simple MPICH cluster in ubuntu. Before , you need an basic knowledge about mpich & clustering. Here we have 4 nodes running ubuntu 7.04 with these host names: ub0,ub1,ub2,ub3;

1. Defining hostnames in etc/hosts/

Edit /etc/hosts like these:

127.0.0.1     localhost
192.168.133.100 ub0
192.168.133.101 ub1
192.168.133.102 ub2
192.168.133.103 ub3

Note that the file shouldn't be like this:

127.0.0.1     localhost
127.0.1.1     ub0
192.168.133.100 ub0
192.168.133.101 ub1
192.168.133.102 ub2
192.168.133.103 ub3

or like this:

127.0.0.1     localhost
127.0.1.1     ub0
192.168.133.101 ub1
192.168.133.102 ub2
192.168.133.103 ub3

otherwise other hosts will try to connect to localhost when they try to reach ub0.

2. Installing NFS

To Install NFS just run this in terminal:

omid@ub0:~$ sudo apt-get install nfs-kernel-server

3. Sharing Master Folder

Make a folder in all nodes, we'll store our data and programs in this folder.

omid@ub0:~$ sudo mkdir /mirror

And then we share it over nodes on the master node.

omid@ub0:~$ sudo echo  /mirror *(rw,sync) >> /etc/exports

Note that we store out data and programs only in master node and other nodes will access them with NFS.

4. Mounting /master in nodes

omid@ub1:~$sudo mount ub0:/mirror /mirror
omid@ub2:~$sudo mount ub0:/mirror /mirror
omid@ub3:~$sudo mount ub0:/mirror /mirror

it's better to change fstab in order to mount it on every boot.

5. Defining a user for running MPI programs

We define a user with same name and same userid in all nodes with a home directory in /mirror. Here we name it "mpiu"! Also we change the owner of /mirror to mpiu:

omid@ub0:~$ sudo chown mpiu /mirror

6. Installing SSH Server

Run this command in all nodes in order to install OpenSSH Server

omid@ub0:~$ sudo apt­-get install openssh­server

7. Setting up SSH with no pass phrase for communication between nodes

First we login with our new user:

omid@ub0:~$ su - mpiu

Then we generate DSA key for mpiu:

mpiu@ub0:~$ ssh­-keygen ­-t dsa

Leave passphrase empty. Next we add this key to authorized keys:

mpiu@ub0:~$ cd .ssh
mpiu@ub0:~/.ssh$ cat id_pub.dsa >> authorized_keys

As the home directory of mpiu in all nodes is the same (/mirror/mpiu) , there is no need to run these commands on all nodes. To test SSH run:

mpiu@ub0:~$ ssh ub1 hostname

It should return remote hostname without asking for passphrase.

8. Installing GCC

Install build-essential package:

mpiu@ub0:~$ sudo apt-get install build-essential

9.Installing Other Compilers

Other prefered compilers should be installed before installing MPICH. In this step we install other compilers such as Inter Fortran, SGI compiler , ... .

10. Installing MPICH2

Download MPICH2 source code from http://www-unix.mcs.anl.gov/mpi/mpich . Extract .tar.bz2 file in /mirror. Also make a folder for MPICH installation.

mpiu@ub3:/mirror$ mkidr mpich2
mpiu@ub3:/mirror$ tar xvf mpich2-­1.0.5p3.tar.gz
mpiu@ub3:/mirror$ cd mpich2­-1.0.5p3
mpiu@ub3:/mirror/mpich2­-1.0.5p3$ ./configure --­prefix=/mirror/mpich2 
mpiu@ub3:/mirror/mpich2­-1.0.5p3$ make
mpiu@ub3:/mirror/mpich2­-1.0.5p3$ sudo make install

For more information about compilation see README file in source package. After successfully compiling and installing mpich, add these lines to "/mirror/mpiu/.bashrc/"

export PATH=/mirror/mpich2/bin:$PATH
export PATH
LD_LIBRARY_PATH="/mirror/mpich2/lib:$LD_LIBRARY_PATH"
export LD_LIBRARY_PATH

Next we run this command in order to define MPICH installation path to SSH.

mpiu@ub0:~$ sudo echo /mirror/mpich2/bin >> /etc/environment

For testing our installation run:

mpiu@ub0:~$  which mpd
mpiu@ub0:~$  which mpiexec
mpiu@ub0:~$  which mpirun

11. setting up MPD

Create mpd.hosts in mpiu's home directory with nodes names:

ub3
ub2
ub1
ub0

and run :

mpiu@ub0:~$ echo secretword=something   >> ~/.mpd.conf
mpiu@ub0:~$ chmod 600 .mpd.conf

To test MPD run above commands. The output should be the current hostname.

mpiu@ub0:~$ mpd &
mpiu@ub0:~$ mpdtrace
mpiu@ub0:~$ mpdallexit

After all run mpd daemon:

mpiu@ub0:~$ mpdboot ­-n 4
mpiu@ub0:~$ mpdtrace

The output should be name of all nodes. If this doesn't succeed try running mpdcheck on all hosts to find possible errors in conf files (they will be marked with "***"). There are some examples in "mpich2-1.0.5/examples", we'll run one :

mpiu@ub0:~$ mpiexec -n 4 cpi

That's it!

---

For more information visit: http://www-unix.mcs.anl.gov/mpi/mpich/