Re: [HTCondor-users] ERROR "Ran out of system resources in auto allocation"; cannot use more than 16 cores on Windows?

[Ralph Finch <ralphmariafinch@xxxxxxxxx>]

Excellent suggestion about the local condor file, I'll do that. It did work with

NUM_SLOTS = ifThenElse(("$(HOSTNAME)" == "DELTA-MOD"), 2, $(DETECTED_CORES))

but as you say, the error from my original method is strange.

Ralph Finch

Calif. DWR

On Wed, Jun 4, 2014 at 3:11 PM, Ben Cotton <ben.cotton@xxxxxxxxxxxxxxxxxx> wrote:

Ralph,

On Tue, Jun 3, 2014 at 6:30 PM, Ralph Finch <ralphmariafinch@xxxxxxxxx> wrote:

> Is there a better way to tell HTCondor that is has fewer cores than actual?
>

What I would do is to use a condor_config.local file on delta-mod that
contains the following line:

NUM_CPUS = 2

(Or if you'd like to future-proof it against hardware upgrades):

NUM_CPUS = $(DETECTED_CORES) / 2

Using a local config file is generally more maintainable than trying
to put host-specific logic in your main config file. When I was at
Purdue, we used a cascade of LOCAL_CONFIG_FILE that defined settings
based on pool, role, and host-specific needs. Depending on the size
and heterogenity of your environment, that might be overkill.
Certainly something like cache_config (which pulls configuration from
CycleServer) or the newly-introduced condor_urlfetch tool might be an
effective way to handle your machine-specific needs.

> I've tracked it down to these two lines in the condor_config:
>
> NUM_SLOTS = $(DETECTED_CORES)
> NUM_SLOTS = ifThenElse(("$(HOSTNAME)" == "delta-mod"), Â2, $(NUM_SLOTS))
>

> Why would the above throw an odd error as previously shown in the log file?
>

I'm not sure why you'd get the error you got, unless HTCondor doesn't
like using the attribute you're setting in an ifThenElse, but even
then I'm not sure why it would bomb out at DETECTED_CORES / 2. If I
were to write the above, I'd do it one line:

NUM_SLOTS = ifThenElse(("$(HOSTNAME)" == "delta-mod"), 2, $(DETECTED_CORES))

Hopefully someone else with better knowledge of the code base can shed
some light on _why_ it behaved the way it did, but I'm glad you were
able to find and work around the issue.

Thanks,
BC

--
Ben Cotton
main: 888.292.5320

Cycle Computing
Leader in Utility HPC Software

http://www.cyclecomputing.com
twitter: @cyclecomputing

_______________________________________________
HTCondor-users mailing list
To unsubscribe, send a message to htcondor-users-request@xxxxxxxxxxx with a
subject: Unsubscribe
You can also unsubscribe by visiting
https://lists.cs.wisc.edu/mailman/listinfo/htcondor-users

The archives can be found at:
https://lists.cs.wisc.edu/archive/htcondor-users/