Re: [HTCondor-users] execute host priority

[Dimitri Maziuk <dmaziuk@xxxxxxxxxxxxx>]

On 02/08/2018 03:20 PM, Todd Tannenbaum wrote:
... RTFM ...

Realistically at this point it's easier for me to shut off the one slow
node with more RAM and lose a couple of cores in the pool, than try to
figure out from the fine manual
- the implications of changing the default NEGOTIATOR_PRE_JOB_RANK
expression, or even just
- what configuration macros I can use there and how, to give faster CPUs
more weight than bigger RAM.

And that is assuming I am correct in that more RAM is what makes jobs
prefer the slow node. That one is least fairly easy to test: just pull
out the extra DIMMs at next scheduled reboot and see what that does.

-- 
Dimitri Maziuk
Programmer/sysadmin
BioMagResBank, UW-Madison -- http://www.bmrb.wisc.edu

Attachment: signature.asc
Description: OpenPGP digital signature