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

Re: [Condor-users] condor pools question



Al,

I saw your post yesterday, but I wasn't certain exactly what you're
trying to accomplish with your function distinctions.  I'll make some
suggestions based on my guess that you're wanting to guarantee that at
least N number of slots is available for jobs of each type you mention.

Let's assume you have all machines pointing at the same central manager:

  CONDOR_HOST = condormanager.ball.com

for instance, set in the main Condor config file, either in a shared
filesystem or copied to all hosts.  All of the settings common to all of
your Condor worker nodes go in that file.

To differentiate between the various sub-pools, you can use local config
files, as defined in the main config file with

  LOCAL_CONFIG_FILE = $(RELEASE_DIR)/etc/local/$(HOSTNAME)

and maybe

  REQUIRE_LOCAL_CONFIG_FILE = TRUE

I would write one config file for R&D, one for Train, one for Test, and
one for Prod.  In the local config directory above, I'd have symlinks
pointing hostnames at those config files.  To shift machines from one
sub-pool to another, you'd switch their symlinks to point to the
appropriate config file and restart the Condor daemons.

In each local config file, you'd use START and/or RANK expressions to
determine what kinds of jobs each sub-pool will run.  You might base
this on the user account or maybe the machine used to submit the jobs.
See this section of the Condor docs:

http://www.cs.wisc.edu/condor/manual/v7.4/3_5Policy_Configuration.html


- dave


On Thu, 2010-08-05 at 10:48 -0400, Berg, Allen wrote:
> We are thinking about using condor, I have a question first
> 
>  
> 
> If I want to create a pool of servers for different functions for
> example R&D, Train, Test, Prod, so in other words I have say 2 servers
> for R&D and 15 in production and want to resize the pool, to have a
> few servers working in training for a class that’s being put on
> tomorrow.   I would just add the servers via the collector_name =
> train.  My understanding is the master node could serve all of the
> different “environments with no issue or further setup.  Is that
> correct?
> 
>  
> 
> Thanks
> 
> 
> 
> This message and any enclosures are intended only for the addressee.  Please  
> notify the sender by email if you are not the intended recipient.  If you are  
> not the intended recipient, you may not use, copy, disclose, or distribute this  
> message or its contents or enclosures to any other person and any such actions  
> may be unlawful.  Ball reserves the right to monitor and review all messages  
> and enclosures sent to or from this email address.
> _______________________________________________
> Condor-users mailing list
> To unsubscribe, send a message to condor-users-request@xxxxxxxxxxx with a
> subject: Unsubscribe
> You can also unsubscribe by visiting
> https://lists.cs.wisc.edu/mailman/listinfo/condor-users
> 
> The archives can be found at:
> https://lists.cs.wisc.edu/archive/condor-users/