Please send an email to Juha or Arto (, ) with
Your name: Affiliation: Email address: Arrival and departure dates: Food allergies, if any:
The meeting room is reserved until Wednesday 31.3., so please indicate in the email if you would like to stay on Wednesday, too. On Wednesday there is also a Grid workshop, you are welcomed to attend that too. Please use the separate registration form from the workshop page.
There are 3 single rooms reserved at the Helsinki university guest house, 40 euros per night, breakfast included. The house is located in down town Helsinki. Please contact Mike for details, .
The closest hotel, within a walking distance from CSC, is Radisson SAS Otaniemi starting from 75/113 euros (single room weekends/weekdays), double rooms 85/113 euros. Must be booked through web 7 days in advance.
In down town area maybe best options are Hostel Academica (single 40-55 euros, double 60-75 euros), Hotel Helka (single 73-101 euros, double 96-129 euros) and Ramada Presidentti (single 84-101 euros, double 91-124 euros).
The bus stops where one should get off the bus (the lines 102 and 103 from the down town Helsinki) are marked on the map. Other possible lines with a slightly different route are 194 and 195. Bus tickets can be bought from the driver (2 euros within a town or 3 euros if you cross town border, like from Helsinki to Espoo) and are valid for about an hour in busses, local trains and trams.
The meeting will be held at TietoEnator building, Tekniikantie 15 a D, Otaniemi, Espoo. Please use the CSC's entrance on the Tekniikantie side of the building. You will get a visitor card from our reception.
Please feel free to contact us for more info:
, +358 50 3819652
, +358 50 5941416
Meeting starts on Monday, March 29, at 10:00 local time.
[Monday 10:00 to 11:00]
[Monday afternoon]
Oxana Smirnova, Aleksandr Konstantinov, Balázs Kónya, Niels Elgaard Larsen, Daniel Kalici, Jakob Nielsen, Henrik Carlsen, Anders Wäänänen, Mario Kadastik, Lauri Anton, Hardi Teder, Andi Hektor, Juha Lento, Mike Gindonis, Jukka Klem, Arto Teräs, Tomas Lindén, Csaba Anderlik, Haakon Riiser, Martin Folkman, Mattias Ellert, Tore Sundqvist, Brian Vinter
In Copenhagen, June 17 and 18.
The production release 0.4 is a week or max. two weeks away. Missing is serious testing, release notes and license. Everybody should please upgrade to the latest version 0.3.38 and test that.
Tag should happen every Friday. If the software doesn't build, Anders will send an email to the mailing list, explaining why a tag doesn't happen.
0.3.38 is available on 12 platforms including SuSE and Redhat 3ES. Work is underway to provide builds on AMD-64.
Anders will setup a contributed-area on the ftp-server.
General complaints about the download-page. It needs a serious reorganization. Anders and Oxana will look at it.
Installation documents SHOULD be maintained! The following persons have volunteered to do the following:
Proper description of middleware comes later.
Balázs has made template-scripts for other LRMS's for the information system. Aleksandr has modified the GM-side submission/information scripts such that there are now 3 scripts. These are the scripts that should be modified to port to another LRMS.
Work is under way for providing a port to SGE.
Haakon presented his nice work on the port to the Condor LRMS. A lot of progress. A prototype seems to work. Only a few things are missing.
Aleksandr has implemented all the features he thinks is necessary for running with afs. He still needs a place to test it though. There might be an afs-storage-element available at DIKU soon.
We still don't know what to do with a Windows Port. Mario will ask a friend of his what one can do.
Mario presented his work on getting the CMS-software to work on NorduGrid. There are lots of problems – especially since the software (like for Atlas) is not available for other platforms than CERN RedHat 7.3. Mario wants to try with a machine running RedHat 7.3 first and get that to run. It should be running within a few weeks.
Jakob can help him with source-rpms for most of the external HEP-software needed when he wants to try on clusters with other distributions.
Mattias presented the Atlas production system with executors, supervisors and the production database. The NorduGrid executor is mostly ready – still needed is RLS-support and testing.
Site-requirements for Atlas DC2: Something like 500 MB ram, outbound internet connection otherwise basically like DC1. Sites should allow a certain subsection of the Atlas VO. This will be specified later.
Total amount of disk-space needed for DC2 is about 100 TB. If NorduGrid does 10%, we'll need 10 TB.
Jakob and Mattias will start building the software for Atlas DC2.
Money for buying 8 clusters. Will probably run the NorduGrid middleware. Details to be clarified. NorduGrid users will probably be allowed to run in some extent.
Henrik presented his nice videorecorder. It works nicely!
Some system-administrators wants to restrict access to runtimeenvironments for certain people – or maybe want to allow only jobs which requests certain runtimeenvironments. Long discussion resulting in nothing.
Martin presented the SGAS. It should be ready for production use next month. It requires GT3 on all clusters.
Niels presented a suggestion for a usage-record. Probably the work should be connected with a survey and comparison of the SweGrid- and GGF-usage record suggestions.
Long discussion about GSIssh. SweGrid are thinking about installing GSIssh on some of their clusters.
Jukka Klem presented the Grid Blocks Portal. You can submit jobs, monitor jobs and so on. Uses the Apache Cocoon framework. Jukka also showed the MonALISA monitoring system which is a centralized information system for (not necessarily grid) systems.
Nothing is to be done at the moment about the XML-RSL. Maybe it would be a nice student-project to study the Globus-XML in GT3.
Jakob has setup an rls-server at rls.nordugrid.org. Everybody has read-access and the developers write-access. Aleksandrs ngcopy and ngremove works with this. We should setup another LRC-server – maybe in Lund. This RLS-server should contain the RC-data.
The Smart Storage Element is basically ready for testing. Everybody is encouraged to try it out.
Daniel has setup two big NDGF Storage Elements. It will run the gacl-plugin so everybody will have access. Missing is still the network connection – should be in place very soon.
The NorduGrid user database will be managed by Anders. The developers and guest-groups will be maintained by Balázs.