[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Condor-users] Easy Tutorial

On Oct 11, 2005, at 5:13 PM, John Alberts wrote:

I am completly new to clustering.  I've read through much of the condor users and admin manuals, but there is a HUGE amount of information.  I have managed to get the basic condor install done on several machines, and it seems to work fine with the examples.
My question is: Now what?  :)  Is there some kind of reference/tutorial/book (Condor for Dummies?) that can help me customize my install for my needs?  Essentially, I plan on having lots of computers in the pool with each possibly having 1 of 3 different programs on them.  How do I identify the computers with a certain program on it, so I can then write a script to submit a job on those computers with the software?

You may find the tutorials from CondorWeek useful: http://www.cs.wisc.edu/condor/CondorWeek2005/presentations.html

For instance, Fluent is one of the programs I would like to know is on a computer in the pool.  The program is usually located in the "c:\Fluent.Inc\ntbin\ntx86" directory.  How would I make a submit file that would identify those machines with fluent on them?

The machines should advertise what programs they have in their machine ads (the ads that condor_status normally displays. To see the full ads, use -l). Then the jobs can require to be matched with a machine which has the program it needs.

For example, put the following in the local config file of the machine:

HasFluent = True

Then, put the following in the job submit file:

requirements = HasFluent =?= True

You can find more information on adding attributes to ads in the manaul: http://www.cs.wisc.edu/condor/manual/v6.7/3_3Configuration.html#10094


|            Jaime Frey            |  Public Split on Whether        |

|        jfrey@xxxxxxxxxxx         |  Bush Is a Divider              |

|  http://www.cs.wisc.edu/~jfrey/  |         -- CNN Scrolling Banner |