Personal tools
You are here: Home Virgo Data Analysis VDAS vdasc_20140324
Log in

Forgot your username or password?
Document Actions


by Gergely Debreczeni last modified 2014-03-24 17:14

VDASC telecon 2014/03/24


  1.  o Summary of last Friday's Dirac deployment discussion (Florent)
  2.  o Short news on LIGO user space software installation plans on XEDE resources (Gergely)
  3.  o Status of Implementation Plan, missing contributions (All)
  4.  o Short introduction of Globus file transfer service and web interface (Gergely) Globus FT
  5.  o Discussion on how, where and who adjoins LIGO frames before sending them to Virgo CC (All)
  6.  o Pegasus testing cluster update (Gergely)
  7.  o AOB


Elena, Mab, Florent, Luca, Lisa, Cristiano, EGO seminar root (Livio, Antonella), Alberto,


1.) Florent presented great progress on Dirac deployment and related discussion for Lyon CC. It seems that current Virgo CC usage can be easily aligned with DIrac framework. More work is needed. Cristiano asked about transparent data access. Answer: Files that are registered in Dirac file catalog can be staged and accessed uniformly independently of the storage backend. This is a good news.

2.) Short news on LIGO's foreseen  user space software installation. On shared resources such as XEDE clusters they will maintain a user space installation of LIGO software. This allows for frequent and more flexible solution compared to the present case, where experiment software is installed system-wide and administration interaction needed to update it. Gergely invited Stuart to on of the VDASC meeting to speak about these plans and the selected tool. It would be nice to synchronize these efforts between LIGO and Virgo, or at least use compatible tools. LIGO is planning to use native relocatable linux packages with "Modules" framewrok. More new later.

3.)Gergely did an ivestigation of the online Globus file transfer service. Short presentation, very convenient tool, does everything with configurable interface and ling-lived proxies. Probably not good / flexible enough for bulk data transfer, but is a very good tool for everyday user wants to do big data movements.

4.) Implementation Plan is progressing slowly, Gergely was asking for more contributions. Florent offered help and asked about the missing parts, which are: some of the data analysis pipelein, comissioning and data acquisition, local job submission for Virgo clusters and for the Cascina online cluster. 

5.) Gergely is progressing with Pegasus / EMI UI testing, software (emu-ui, condor, pegasus) installed, configured. Local job submission works finem promised more new and small tutorials for next week.

6.) Long discussion on data transfer. Agreement on that it should be a Virgo member repsonsible for merging and transfering LIGO frames to Virgo CCs. Would be nice to invite Stuart to speak about experiences, since LIGO folks were doing this berfore. There are some question to clarify: a.) Why there is a need for 1 day maximum allowed latency, b.) is it possible to have a smaller set (duration) of LIGO files, so the timer ordering problem, does not requires so much extra space c.) there are bandwith limitations which could result accumulation of data, need to be clarified. We will continue this next week.