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

Re: [Condor-users] Jobs stay Idle ... been looking for 24 hours....



On Thu April 26 2007 1:46 pm, Askar Zaidi wrote:
> Hi,
Hello,

> My jobs stay idle forever...
Yup.  ;)  See below.


> vm1@xxxxxxxxx LINUX       INTEL  Owner      Idle       0.000   378
> 0+00:10:09
<snip>
> vm3@xxxxxxxxx LINUX       INTEL  Owner      Idle       0.000   504
> 0+00:10:11
> vm1@xxxxxxxxx LINUX       X86_64 Owner      Idle       0.890   250
<snip>
> vm4@clouseau. LINUX       X86_64 Unclaimed  Idle       0.000   250
> 0+00:10:07
> vm1@dogmatix. LINUX       X86_64 Owner      Idle       0.110   501
> 0+00:10:10
<snip>
>
>                      Total Owner Claimed Unclaimed Matched Preempting
> Backfill
>
>          INTEL/LINUX     8     8       0         0       0          0
> 0
>         X86_64/LINUX    16     9       0         7       0          0
> 0
>
>                Total    24    17       0         7       0          0
> 0


> ( target.Arch == "INTEL" ) && ( target.OpSys == "LINUX" ) &&

If you look at all of the above, you'll see:

1. Your job's requirements specify 'Arch == "INTEL"'
2. While there are indeed some INTEL/LINUX machines in the pool, they are all 
in the "Owner" state, and thus won't run jobs.  The others are all "X86_64", 
which (as far as Condor is concerned) different that "INTEL".

So, what should you do?

1. Look at the START expression on the 32 bit, see why they're all in the 
Owner state.
2. Allow your job to run on "INTEL" and "X86_64" machines (after all, these 
are "INTEL" binaries, and should run on either IA32 or X86_64, right?).  You 
can do this by adding something like this to your requirements expression:

Requirements = (  (OpSys == "LINUX") && ((Arch == "INTEL") || (Arch 
== "X86_64" ))  )

This will for Condor to match your job to machines with the Arch attribute 
as "INTEL" or "X86_64".

Hope this helps

-Nick

-- 
           <<< There is no spoon. >>>
 /`-_    Nicholas R. LeRoy               The Condor Project
{     }/ http://www.cs.wisc.edu/~nleroy  http://www.cs.wisc.edu/condor
 \    /  nleroy@xxxxxxxxxxx              The University of Wisconsin
 |_*_|   608-265-5761                    Department of Computer Sciences