NUC 09 Oct 2013

Attendance

Andreas Haupt; Yves Kemp; Kai Leffhalm; Shaojun Lu; Christoph Rosemann; David South; Hartmut Stadie; Michael Steder

Agenda

https://indico.desy.de/conferenceDisplay.py?confId=8909

News

HGF alliance meeting Dec 2-4
looking for a new chair, Dave will be chair starting in January

NAF status

NAF-2 migration:
- NAF 2.0 basically ready (registration, services);NAF-1 should be shut down in spring 2014
- AFS:
  - home directories need to copied over from naf.desy.de to desy.de cell
  - additional space via xxl volume(no backup, cheaper hardware)
  - access: kinit foobar@DESY.DE; aklog -c desy.de
  - group space done by experiments
- dCache: unchanged, dcTools will be replaced by direct read-only dCache mount on SL6 wgs
- Lustre decommissioned since Oct 1st
- SONAS:
  - a directory cannot be mounted in parallel at NAF 1 and 2, migration can be done per user directory (fileset)
  - status: ILC migration finished, ATLAS& CMS only a few users migrated
  - capacity: 650 TB, 200 TB purchased
  - user directory should have group id of external users (e.g. af-cms), have to change group at login for internal users (mechanism working for ATLAS and ILC, but needs work for CMS)
- workgroup servers:
  - users can work in parallel on NAF 1 and 2
  - NAF 2.0 wgs ready on SL5, additional servers with SL6 being deployed
- batch:
  - NAF 2.0 1600 cores with SL5 and 1000 cores with SL6
  - more resources to come: 400 + 400 cores SL6
  - migration from NAF 1 to 2: Oct 500 cores, Nov 1000 cores, Dec 1000 cores, Jan 350 cores, March 150 cores, April shut-down
  - Zeuthen will more its 120 cores to local batch farm or decommission
- others:
  - monitoring of all components: some tools exist, batch monitoring to come
  - documentation: http://it.desy.de/services/computing_infrastructure/national_analysis_facility___naf/index_eng.html
  - fine-tuning of WGS rpms, etc
questions:

Open action items

id	who	description	status/comments
1007-2	NAF	report on help desk tickets	shown
1210-2	NAF 2.0	offer better monitoring for queue usage (per user, waiting time, run time, etc)	will come
1210-3	NAF 2.0 & experiments	reconsider queue setup, scheduling, dynamic or configurable queue limits	not yet
1302-2	NAF admins	clearly document what scratch space means and policy for Lustre and SONAS	Lustre done

ATLAS report

cvmfs mount missing on ATLAS wgs -> mail unix@desy.de
mailing list setup will reconsidered

CMS report

ROOT crashing on NAF-2 wgs? mail: unix@desy.de

ILC/CALICE report

works smooth on NAF 2.0
problems with different OS versions and new gcc versions with new C++ standard

LHCb report

BELLE report

needs local mount to read data with BELLE-1 software

AOB

New Action Items:

who

description

status/comments

Next meetings

Wednesday, November 13th 2013, 1 pm, https://indico.desy.de/conferenceDisplay.py?confId=9214