Follow @Openwall on Twitter for new release announcements and other news
[<prev] [next>] [day] [month] [year] [list]
Date: Tue, 21 Dec 2010 00:49:45 -0600
From: definitely crashing <slowlydryingup@...il.com>
To: john-users@...ts.openwall.com
Subject: Attempting to run John on Amazon EC2 Using SGE

I'm fairly new to all of this (Linux, EC2, StarCluster, SGE, etc...) so
stick with me if this is incredibly dumb.

I've successfully installed and figured out how to run john on my local
cluster using openmpi.  The next step was that I decided I should try to
scale up even further to crack using Amazons EC2 instances.  I was able to
get a cluster up and running and it seems to respond and exeute my runmpi
commands sueccessfully.  SGE does a great job of managing the "slots" across
multiple instances.

I have compiled john 1.7.6 w/ both the Jumbo 8 patch and the FullMPI patches
listed on the wiki.

Everything runs fine but no matter what I do, I can't get john to run
multiple sessions on the amazon instances.  SGE executes the mpirun command
successfully but it never shows john as using more than 1 slot/core.  I can
specify the -np x option, but SGE recommends not doing this (it doesn't
matter, anyway) and it is supposed to handle all of the openmpi options on
its own.  Is there some way I can force john to run multiple instances?  I
believe it may not be detecting the openmpi environment when it is invoked
by SGE?  Anyone figure out a way to successfully run it with mpi patches on
EC2?

Powered by blists - more mailing lists

Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.