NAF User Committee Monthly Meeting
→
Europe/Berlin
7/7a (DESY Hamburg)
7/7a
DESY Hamburg
Minutes of NUC meeting from October 12th, 2009
Present/Phone: Johan Blouw, Alexey Zhelezov (LHCb),
Jan Erik Sundermann(ATLAS), Hartmut Stadie(CMS),
Steven Aplin, Niels Meyer, Angela Lucaci-Timoce(ILC)
Andreas Haupt(IT)
1. News from chair:
We welcome a new member Angela Lucaci-Timoce(ILC) who will replace Niels at the end of this year.
2. Action items:
i) documentation and user information:
-NUC members welcome to give feedback to NAF admins on the new
news page at
http://naf.desy.de/general_naf_docu/faq_and_support/news/
ii) update on cmt compile problem:
-some issues fixed, ongoing
iii) NAF SL5 migration:
-WN will stay at 10% SL4, 90% SL5 til next meeting;
-ATLAS and CMS have to further test SL5 WGS, ILC and LHC-B are
ready for migration of default WGS to SL5.
iv) different sysname for SL5 compared to SL4:
-will be discussed offline between ILC and the admins.
v) important topics for NUC face-to-face meeting:
-detailed presenation on the SGE batch system(fair share, how
the queues work...)
-use cases for storage systems, what is needed?
vi) advertise NAF user meeting at HGF workshop:
-please do so.
vii) where are SL4/SL5 nodes located(hh/zn):
-can be seen with qhost command.
- remaining SL4 nodes equally distributed between HH and ZN
-> action item closed
viii) present monitoring of WGS:
-history information can be seen using the "sar" command on WGS.
Please give feedback.
-publication of central monitoring complicated as they
also contain security relevant information
ix) advice on user files storage:
-will be discussed next meeting
-admins will update storage documentation on the web
3. status report:
please see the slides from Andreas:
- serious problems with HH lustre instance: -> new action item
- several crashes affecting HH instance of lustre
- complete shutdown recovered the instance, but problem not fully
understood.
- as a temporary fix(order half a year) providers will setup a new instance and ask users to migrate their files to the new instance.
- the procedures will be sent to all users via naf-announce.
- providers investigating other storage solutions
- switch from SL 5.3 to SL 5.4 as the SL5 operating system.
- will be tested on WGS first. No objection from NUC to move to SL5.4
on the existing SL5 WGS
4. AOB:
- a reminder that, every experiment should briefly present one case of
their NAF usage at the HGF NUC session
- the NUC thanks the providers for quickly coming up with a solid plan
to solve the current stability problems with the HH lustre instance
- next NUC is face-to-face meeing in Hamburg on November 11th, 10:00
There are minutes attached to this event.
Show them.