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

Re: [HTCondor-users] SYSTEM_JOB_MACHINE_ATTRS



Hi Steven,

On 21/01/2019 14:37, Steven C Timm wrote:

MachineRalScaling is not a default HTCondor attribute, it must be something that is being defined in GridPP clusters somehow.

It's defined for our APEL accounting. We attach it to "the job" via SUBMIT_EXPRS on the head node.

MachineRalScaling = "$$([ifThenElse(isUndefined(RalScaling), 1.00, RalScaling)])"
SUBMIT_EXPRS = $(SUBMIT_EXPRS) MachineRalScaling

Since the $$ syntax is used, expansion is delayed until the job gets on the worker-node. In the worker node, a local value (RalScaling) is substituted in. This gives the power of the node, hence we can have heterogeneous worker-nodes and the the power expression comes out in the job data. The MachineRalScaling "emerges" in the condor_history data. It's then a simple matter to multiply the wallclocktime by  MachineRalScaling to "normalise" the job, i.e. make them all the same.

# condor_history -long 1233764.0 | grep MATCH_EXP
MATCH_EXP_MachineRalScaling = "1.036000000000000E+00"

This has stopped happening, in the new condor I use. I was wondering if the behaviour has been changed, somehow. I got it working with SYSTEM_JOB_MACHINE_ATTRS.... but I'd like to keep things the same.  I have a feeling that MATCH_EXP_* is an "undocumented" feature, since I can't see it anywhere in the manuals.

Cheers,

Ste


--
Steve Jones                             sjones@xxxxxxxxxxxxxxxx
Grid System Administrator               office: 220
High Energy Physics Division            tel (int): 43396
Oliver Lodge Laboratory                 tel (ext): +44 (0)151 794 3396
University of Liverpool                 http://www.liv.ac.uk/physics/hep/