I don't know what the problem in this case is, but e.g. we had lots of
problems with CMS sending thousands jobs (often in one go) to the CE
with the lowest response time, despite the fact that we have 4 CEs
fronting the same cluster and the response times would only differ
marginally. We are now rounding the response times we send out and the
load gets much more evenly distributed. It's easier to do this, than
to convince various bit of software that the last digit or so really
isn't relevant.
Daniela
On 12 July 2012 13:09, Jeff Templon <[log in to unmask]> wrote:
> Hi
>
> what do you mean that you want to edit it? what is wrong with the value printed?
>
> JT
>
> On Jul 12, 2012, at 13:55 , Maarten van Ingen wrote:
>
>> In this case I want to edit the "estimated response time" which is already published by "glite-info-dynamic-scheduler-wrapper".
>> I could edit this file, but then I have to every time i run YAIM and I do not wish to edit/hack files modified by YAIM.
>>
>> Cheers,
>> Maarten
>>
>> On Jul 12, 2012, at 11:59 AM, Stephen Burke wrote:
>>
>>> LHC Computer Grid - Rollout [mailto:[log in to unmask]]
>>>> On Behalf Of Maarten van Ingen said:
>>>> What are the conventions and guidelines when building plugins for the
>>>> (resource)bdii?
>>>
>>> What are you trying to do? Normally the order of the plugins doesn't matter because each plugin deals with a different LDAP object so they don't interfere. If you want to change the behaviour of an existing plugin it would be better to edit/replace it.
>>>
>>> Stephen
>>>
>>> --
>>> Scanned by iCritical.
>>
--
-----------------------------------------------------------
[log in to unmask]
HEP Group/Physics Dep
Imperial College
Tel: +44-(0)20-75947810
http://www.hep.ph.ic.ac.uk/~dbauer/
|