TMB - 2008-07-16
Present: Oliver Keeble (chair), Markus Schulz, John White,
Francesco Giacomini
Phone: Vangelis Floros, Cal Loomis, Claudio Vuerli (Astrophysics
cluster), Mariusz Sterzel (Computational Chemistry cluster),
Francisco Castejon (Fusion cluster)
Apologies from Jeff Templon and Andrei Tsaregorodtsev
Minutes of last meeting
=======================
No comments.
Action list
===========
We skip the task with priority set to "Later".
Task #5401 (Identify resources for LSF plugin support)
Oliver: no resources for this in sa3
Task #6337 (SA3 priority list)
Oliver: I have update the task, waiting agreement with Erwin that it
is sufficient
Task #6649 (Local scratch space and shared storage in Glue)
This comes from NA4 requirements list; NA4 to check the latest post;
please do it soon, since GLUE2 is now in public comment.
Assign the task to Cal.
Task #6651 (Examine requirements for publishing java and python versions)
No news.
Task #6652 (Check if the WNRWG will make recommendations about how to
characterise a subcluster)
This task depends on #6650, which has originated bug #37911 (ISM
purchaser should handle Glue 1.3 subcluster software entity). If that
is sufficient, Steve can start putttin together his recommendations.
Task #6718 (check status on gLite delegation library)
John: it's done, no replies, close it
Task #6823 (Write a proposal on multi-platform support and circulate to ROCs)
No news, no site representatives at the meeting.
Task #6901 (Storage semantics issues for EGEE (beyond HEP)
Vangelis: the task was assigned to someone in INFN, a draft should be
available soon.
Task #7038 (Check that the proxy chain verification is done correctly
in the current services)
John: all information received, close it
Task #7106 (identify the Glue attributes which require the definition
of a common policy on their meaning)
No news, no SA1 representatives at the meeting.
Task #7160 (collect info on R-GMA curren use)
Francesco: information not yet available
Oliver: the information should include not only what are the uses of
RGMA, but also who relies on it
An item on RGMA on the agenda next time.
Task #7162 (collect inforamtion on R-GMA usasge)
Cal: only the dashboard uses RGMA as far as I know, and they can move
away from it; I'll circulate another mail just to be sure.
Task #7163 (report xrootd security issues to dCache)
John: we have a meeting coming up; I will report
Task #7187 (end use case to GridObservatory)
No news, no SA1 representatives at the meeting
Task #7188 (forward intersting use case to GridObservatory)
Seems identical to the previous one, both are assigned to SA1. Assign
to Francesco, to check if this in fact is for JRA1.
Task #7281 (Schedule a full discussion on pilot job framework at a future TMB)
Postpone the discussion to a next meeting
Task #7282 (Markus and Oliver to discuss with Alain Roy about bringing
OSG and gLite UIs closer)
Oliver: we discussed it with Alain last week: the simplest solution is
that OSG takes some bits of gLite and builds them; they cannot take
directly binaries because there are differences in dependecies,
e.g. globus. In any case the discussion will not result in a common
UI.
Oliver to comment in the task and close it.
NA4 workplans
=============
Cal: we will go through the attached document and describe the tasks
for three clusters: Astrophysics (with Claudio Vuerli), Computational
Chemistry (with Mariusz Sterzel), Fusion (Francisco Castejon)
TMB more interested in the parts of the workplan that concern
development.
*** Astrophysics
Discussion/emphasis on the following tasks:
AA3 (Applications)
Claudio: the task is about application porting to the Grid. The plan
is to strenghten the contacts with developers to speed up and drive
the activity. Reuse Grid tools and services already available; we
devolop our own as last resort. Access to databases on the Grid
is important; there is GRelC in the RESPECT program, another database
access tool was developed with INFN outside of egee, we may
port this in EGEE. We will bring requirements to the TMB for further
needs from middleware.
AA5 (Interactions with EuroVO)
Claudio: the interaction with the European Virtual Observatory is
strategic for our activity. This task represents a cross-link with a
similar workpackage within that project.
Workplan approved by TMB.
*** Computational Chemistry
CC4 (Extension of Charon GUI)
Mariusz: Charon can be used by any community, it provides an
abstraction of job management on the Grid
The TMB suggests to specify what original features it has.
CC5 (Extension of Wien2K Workflow)
Mariusz: main focus is on reducing overhead for short jobs
CC7 (Development of Grid license models)
Mariusz: licensing is an important issue
The TMB suggests to check how this information can be published with
the GLUE schema, especially with GLUE2.
CC9 (Port of parallel versions of chemical software packages)
Mariusz: we have the need for more MPI-enabled resources
Oliver: good news: there is an SA3 partner (from Russia) interested in
MPI; they can follow the configuration issues
Workplan approved by TMB
*** Fusion
F4 (Data Management)
Markus: are EGEE tools used?
Francisco: yes, if suitable
Markus: what are the storage suppliers?
Francisco: some labs involved in the work
TMB suggests to include a first milestone on requirements and
assessment of existing tools. Francisco proposes also a milestone with
a final report.
F5 (Workflow)
Oliver: consider the interoperability work done by others within the
project, SA3 in particular, to be able to run jobs on other
infrastructure
Workplan approved by TMB
*** Other clusters
The workplans for the other clusters will be discussed the next
meeting(s).
AOB
===
1. there is a new operations requirement from Romain, concerning the
collection of some security information on the gLite services. This
is similar to something SA3 is trying to finish (a reference card
for each service) and similar to what the monitoring group does. It
is agreed that this information should be included in the
documentation of each service. This will be included in the
acceptance criteria set by SA3.
2. JRA1 requests that requirements are kept in savannah and not in a
twiki, for easier reference from other savannah trackers. To be
discussed at a next TMB.
There are minutes attached to this event.
Show them.