Hi,
thanks. Bad package name. I will calm down now. Would be really great to rename the package. It does nothing but report dynamic LRMS info. No scheduling.
Bad package name. One could even just remove "scheduler" (and one hyphen) and it would be ok ... glite-info-dynamic-lsf is just fine.
But not scheduler.
J "did I say bad package name?" T
ps: back to Matt: your mail seems to indicate that somehow the config got messed up, as indeed the line you added to lrms_backend_cmd needs to be present.
The other place to look, take a good critical look at the vomap section of the scheduler conf file ... YAIM gets this wrong once in awhile (are you using YAIM?)
On 2 Dec 2010, at 17:07, Massimo Sgaravatto - INFN Padova wrote:
> Mmm ?
>
> glite-LSF_utils provides both:
>
> glite-info-dynamic-scheduler-lsf
> lcg-info-dynamic-scheduler-generic
>
> The first one is the one provided and maintained by Cern
> (Ulrich Schwickerath)
>
> It provides:
>
> /opt/glite
> /opt/glite/etc/glite-info-dynamic-lsf.conf
> /opt/glite/lib/perl5/CommandProxyTools.pm
> /opt/glite/libexec
> /opt/glite/libexec/glite-info-dynamic-lsf
> /opt/glite/libexec/lrmsinfo-lsf
> /opt/glite/share/doc/glite-info-dynamic-lsf.txt
>
>
>
> Cheers, Massimo
>
> On Thu, 2 Dec 2010, Jeff Templon wrote:
>
>> Holy Moley,
>>
>> On 2 Dec 2010, at 16:33, Matt Doidge wrote:
>>
>>> rpm -qa | grep lsf
>>> #glite-apel-lsf-2.0.10-1
>>> lsf-7.0.6-1
>>> lsf-lmgrd-7.0.6-1
>>> glite-info-dynamic-scheduler-lsf-2.0.7-1
>>
>> is there also an lcg-info-dynamic-scheduler installed????? Will you do an rpm -qa on the glite-info-dynamic-scheduler-lsf package? I really hate this, finding out I am trying to debug either a hacked version of my plugin, or else a horribly named but otherwise valid package.
>>
>> Plugins for the dynamic-scheduler should be called something like
>>
>> lrmsinfo-lsf
>>
>> because they are not making any scheduling predictions themselves, they are just dumping INFO about the LRMS state for LSF ...
>>
>> There should *never* be a glite-info-dynamic-scheduler-lsf (or -anything-else) except for -generic. If it only provides the lrmsinfo, it should be named as above. If it really is computing scheduling predictions, it should be uninstalled.
>>
>> This is by design, to avoid biases between LRMS flavors, the algorithm is completely statistics based and hence exactly the same for all LRMS flavors.
>>
>> JT
>>
>> ps: Matt : this rant is not for you, sorry.
>>
>
> \|||/
> -----------0oo----( o o )----oo0-------------------
> (_)
> INFN Sezione di Padova
> Via Marzolo, 8
> 35131 Padova - Italy E-mail: massimo.sgaravatto [at] pd.infn.it
> Tel: ++39 0498275908 Skype: massimo.sgaravatto
> Fax: ++39 0498275952
|