And the problem might not be the CE at all. For example Manchester had a
very bad ranking with lhcb because of stale bdii cache and for the same
reason biomed submitted 12k jobs two days later.... in short you need to
know your VOs patterns. But if it is nagios you are worried about it
only checks your CE is in the BDII and in production.
cheers
alessandra
On 20/05/2011 12:52, John Gordon wrote:
> I think this is a special case of the more general problem - 'Jobs have stopped coming to my site from VO=xxxxx' It may be something you have done that has made your site less attractive to the matchmaking or it may be you have been blacklisted.
>
> John
>
>> -----Original Message-----
>> From: Testbed Support for GridPP member institutes [mailto:TB-
>> [log in to unmask]] On Behalf Of Sam Skipsey
>> Sent: 20 May 2011 12:27
>> To: [log in to unmask]
>> Subject: Re: BrokerHelper: no compatible resources
>>
>> On 20 May 2011 12:20, Stephen Jones<[log in to unmask]> wrote:
>>> Stephen Burke wrote:
>>>> Whether it would be useful is another matter - you would hardly want
>> the
>>>> details of matches against every CE in the BDII ...
>>>>
>>> Yes - that's a lot of data just to tell me the resource my CE lacks.
>>>
>> But, again, I'm missing the critical importance of this to your site
>> functioning correctly.
>> If *all jobs to the WMS are failing to match against your CE*, then
>> obviously it's a problem that your CE isn't publishing correctly.
>> Other than the two cases of: CE is not publishing at all (easily
>> picked up by the BDII tests we already have for a CE service), and
>> Queues are closed at site (which you should be aware of at your site),
>> jobs generically failing to match against your CE aren't a problem for
>> the site (they just indicate that the user wanted things that aren't
>> at your site - but that can be software only installed at one
>> location, data only in 3 places in the world, the name of a CE being
>> explicitly chosen, etc, all things that aren't a problem with your
>> site, since the users' job still runs at a site that does provide
>> them!).
>> Is this genuinely an issue you've encountered, or just a hypothetical
>> problem?
>>
>> Sam
>>> Steve
>>>
>>> --
>>> Steve Jones [log in to unmask]
>>> System Administrator office: 220
>>> High Energy Physics Division tel (int): 42334
>>> Oliver Lodge Laboratory tel (ext): +44 (0)151 794
>> 2334
>>> University of Liverpool
>> http://www.liv.ac.uk/physics/hep/
|