[Hps-analysis] tweak pass6

[Rafayel Paremuzyan]

Hi all,

after about 1 day of running of farms, most of the jobs failed,
it appeared to be a global JLab wise problem with the lustre system.

Currently they are working on resolving it.
At this moment roughly 15% of data is processed.
I will update, if I get some more information

Below is a responce of a ccpr.

Rafo

The I/O errors are due to filesystem issues with Lustre -- several of
the storage targets became unavailable and will not remount. We are
working the issue with Intel ...

Sandy




On 12/30/2016 11:17 PM, Rafayel Paremuzyan wrote:
> Hi all,
>
> Today tweakpass6 has started.
> Thanks to Holly and Omar who are main contributors to this pass.
>
> In addition to tweaking tracking parameters, this pass also includes
> -Track states in 1st and 2nd layer of SVT is added
> -Event time is calculated using the RF time, which is determined 
> through the highest energy cluster in the event. The cluster indices 
> of that crystal is preserved
> -This pass has a new format root files "from ntuple maker", that were 
> previously used, but wasn't considered in the pass, it includes 
> Moller, fee, and tri skims
>
> Given the fact that it was hard to make a release during the Christmas 
> New year brake,
> and data is need soon for bump hunt publication, we didn't have a 
> release, but just have used a tested revision 4657.
>
> Data can be found on tape:
> /mss/hallb/hps/engrun2015/tweakpass6/
>
> Confluence page for this pass is
> https://confluence.slac.stanford.edu/display/hpsg/Tweak+pass+6
>
> Rafo