[mpich-discuss] Mpich2 - in Red Hat 7.3

[Sandra Guija]

Hi Gus,The reason I'm interested to have a local copy in each node is because I am working to compare analysis performance between message passing and shared memory programming model. I am trying to avoid NFS delays against the message passing model.
One question I have regarding your email is when you said: "hassle of copying and maintaining exact replicas on the compute nodes"I just setting up an lab environment, not to be connected to a real network, I am not planning to do any upgrades, so what will be the reason the replicas will not be the same.Thanks,Sandra Guija

> Date: Fri, 5 Oct 2012 18:13:12 -0400
> From: gus at ldeo.columbia.edu
> To: mpich-discuss at mcs.anl.gov
> Subject: Re: [mpich-discuss] Mpich2 - in Red Hat 7.3
> 
> On 10/05/2012 03:35 PM, Pavan Balaji wrote:
> >
> > On 10/05/2012 02:33 PM, Sandra Guija wrote:
> >> If I decide to setup my cluster so that each node has their own mpich2
> >> binaries,
> >> it will be ok if I installed just in one and the copied the
> >> mpich2-install directory to the other node?
> >> Or I need to install idnivdually
> >
> > It's theoretically OK to just copy the install directory, but you have
> > to be very careful to maintain correct permissions, and directory paths,
> > etc. This is often a cause of many errors.
> >
> > -- Pavan
> >
> 
> Hi Sandra
> 
> As Pavan said, you can copy over the install directory to the
> nodes, with caution to the details that Pavan pointed out.
> 
> Also, make sure all computers have the same architecture also,
> i.e., say all 32-bits (x86), or all 64-bits (x86_64).
> 
> However, I believe it is much simpler for you to install mpich2 only
> in the cluster's head node, on a NFS (network file system)
> shared directory.
> I.e., install in a directory on the
> head node that is exported to the compute nodes
> and mounted on the compute nodes.
> Typically people create a directory only for this purpose,
> called, say, /shared, and you could install mpich2,
> say, in /shared/mpich2 or similar.
> 
> The head node is the NFS server, the compute nodes are
> the NFS clients.
> 
> This will require a single install of mpich2.
> You will avoid the hassle of copying and maintaining
> exact replicas on the compute nodes.
> In small clusters this is typically what people do.
> 
> NFS is not much of an mpich2 list issue, but anyways ...
> First check if you have nfs packages (if you're on Fedora use yum)
> 'yum list |grep nfs'
> and install the packages with 'yum install' if you don't have them.
> Second if you already have NFS packages, check if they're
> running on the head node:
> 'service nfs status'.
> If NFS is not running, try
> 'chkconfig --list |grep nfs'
> If NFS is turned off, turn it on, and start the service:
> 'chkconfig nfs on'
> 'service nfs start'
> 
> If you are not familiar to NFS,
> and how to export and mount a directory,
> here is the Linux NFS 'HowTo':
> 
> http://nfs.sourceforge.net/nfs-howto/
> 
> I hope this helps,
> Gus Correa
> _______________________________________________
> mpich-discuss mailing list     mpich-discuss at mcs.anl.gov
> To manage subscription options or unsubscribe:
> https://lists.mcs.anl.gov/mailman/listinfo/mpich-discuss
 		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20121005/7e8d36a0/attachment.html>