Dear All,
Please find attached the weekly GridPP Project Management
Board Meeting minutes. The latest minutes can be found each week in:
http://www.gridpp.ac.uk/php/pmb/minutes.php?latest
as well as being listed with other minutes at
http://www.gridpp.ac.uk/php/pmb/minutes.php
Cheers, Tony
PS Note that the minutes refer to this Friday 29th April as the deadline
for the LCG 2.4.0 upgrade at UK sites.
________________________________________________________________________
Tony Doyle, GridPP Project Leader Telephone: +44-141-330 5899
Rm 478, Kelvin Building Telefax: +44-141-330 5881
Dept of Physics and Astronomy EMail: [log in to unmask]
University of Glasgow Web: http://ppewww.ph.gla.ac.uk/~doyle
G12 8QQ, UK Video - IP: 194.36.1.32
________________________________________________________________________
GridPP PMB Minutes 169 - 25th April 2005
========================================
Present: Tony Doyle, Steve Lloyd, Robin Middleton & Tony Cass by video.
Dave Britton & Sarah Pearce by phone and Skype.
Apologies: Jeremy Coles, Dave Kelsey, Roger Jones, John Gordon & Peter Clarke
Steve Fisher in attendance for first agenda item.
1. R-GMA status and plans
==========================
SMF was asked to attend this meeting to give an update on the current
status of R-GMA and plans for implementation. He started by looking at
some basic observations: He said there appeared to be different opinions
on how difficult R-GMA was to deploy. This was partly due to a problem
with gLite introduction of authentication. A change of port was
required, but as people were failing to read instructions, this was not
being understood.
It was no longer true that R-GMA has a single point of failure in the
registry. The Schema replication still has to be coded however, so that
everyone sees the same authorization rules. Regarding reliability,
there had been some problems with R-GMA multiple processors. This now
appears to be fixed with the latest JDK. He also said that applications
must catch exceptions as R-GMA won't work if they are ignored. He
disputed the criticism that R-GMA was not being used but agreed that
that there was a problem with large history queries.
SMF went on to report on the meeting with LCG last week who are now
making extensive use of R-GMA. He said LCG saw no advantage to
switching to the gLite version at the moment. The task is to make the
gLite version compatible with LCG2_4_0 so that it can be deployed in the
usual way. Returning to the concerns about history queries, this was due
to the archiver running out of memory. This is because when a query is
issued it sits in the server which returns too many tuples which then
crashes. LCG also requested that the first tuples should start being
returned quickly. This all needs careful planning.
TD asked about authentication not being required by LCG and SMF said Ian
Bird preferred no authentication as it stops them doing a simple upgrade.
TD referred to the problem with the server crashing and said he would
provide a URL on metadata slides relating to back-end databases which
is being used for Mon-Wednesday's metadata workshop in Grenoble
http://lpsc.in2p3.fr/cdsagenda/fullAgenda.php?ida=a0557%20
Looking at Deployment issues, SMF said there had been a problem with the
2_4_0 release and said the old tomcat RPM would have to be removed. This
is not a problem with clean installs but only with upgrades. TD asked
who the contacts were who could address deployment current concerns.
SMF said most problems were dealt with by Steve Traylen or Laurence
Field from monitoring the ROLLOUT list. Otherwise the R-GMA team pick up
the problems. Although there were good instructions on the web pages,
things often did not make it to the LCG instructions. Weekly meetings
with LCG and R-GMA were being planned however to discuss issues.
LCG have modified the job wrapper to publish a tuple to R-GMA from every
job every 5 minutes. In the LCG2_4_0 other information including from
L&B and APEL is being published. The component to populate the
"Information Supermarket" from R-GMA is being worked on in Milano. This
will eliminate dependency upon the BDII.
SMF said talks had taken place with DILIGENT about 'end user'
applications and Bioinformatics had shown some interest. Unfortunately
HEP experiments preferred, in general, to go it alone.
SMF said they had considered using the OGSA-DAI DQP mechanism but it would
not be useful at present. This is because it can't be integrated easily
as it's not Java and basic DQP techniques don't work properly with R-GMA
because of duplicate tuples existing in the system.
SMF noted that the 15th July web service release will be delayed, with an
as yet undefined release date, because of the extra work for LCG to fix
the large History queries. APIs have already been released. Concerning
metadata catalogues, R-GMA may be applicable as jobs could publish tuples
appropriate to various searches.
TD said he would send SMF two documents where R-GMA could help:
http://ppewww.ph.gla.ac.uk/~mlyon/HEPMetadataSchema_v0_7.pdf
http://www.gridpp.ac.uk/datamanagement/metadata/SubGroups/UseCases/
In summing up SMF said everyone in the R_GMA team was working hard on the
problems as they came up. TD said promoting LCG was good but a plan was
needed for the web service release delay. SMF said more discussion was
needed with LCG and the rest of JRA1 to come up with a plan. TD said the
LCG side would benefit more if interaction was more formalised. SMF said
this should be addressed by the new weekly meetings between JRA1-UK and
LCG.
It was agreed that SMF had highlighted all the areas of criticism and
verified that these were being addressed. He was thanked for attending
today's meeting and RM suggested having an update at a point when it was
suitable.
2. EGEE REPORT FROM ATHENS
===========================
RM said the conference had gone very well with positive feedback from
the External Advisory Committee. Among the concerns is the deployment of
gLite. Significant effort is now being given to preparing the proposal for
phase 2 of the project.
3. QUARTERLY REPORTS
=====================
It was planned that DB would have these completed within the next 2 weeks.
All area managers would need to provide their input and a summary prior
to this.
ACTIONS
=======
157.2: RJ - review experiment computing models for Tier-1 to Tier-2
interactions in the UK.
ongoing
157.9: RM to report on direction of GridPP in terms of middleware.
ongoing - gLite is now out so this can be drafted
157.12: SP to provide standard template talk for schools.
ongoing - BP is still working on this at present
159.11: RJ to make Andrew fully aware of ATLAS plans
ongoing
159.13: TC to extract LCG issues from GridPP12 talks
incomplete - (change to provide input for GridPP13 talks)
159.14: RM to extract Middleware issues from GridPP12 talks
ongoing - awaiting info
159.19: DB to make a list of all GridPP people and their roles
ongoing - DB has received some comments on the draft. SP will send an
additional spreadsheet against which to cross-check names
168.1: RM to post all quarterly report logbooks to a webspace
done - AMcN set it up last Friday
168.2: JC & AS to discuss and finalise production grid milestones &
metrics for Tier-1
ongoing
168.3: SP to liase with DB to update GridPP people and their roles for
the spreadsheet being prepared
done
ACTIONS FROM FACE TO FACE MEETING
=================================
167.1: JG to report back from networking meeting on service
challenge activity/organisation.
ongoing
167.2: JC, with input from Stephen Burke to maintain a list of middleware
issues that should be addressed
ongoing
167.3: RJ to define UK contacts for each experiment who will
respond on experiment operational issues.
ongoing
167.4 TD to raise with RM possibilities for funding HEP SysMan
meetings
ongoing - but changed to 'discuss revised funding policy for GridPP people
attending HEP SysMan meetings'
167.5 DK to define a possible technical documentation role.
ongoing
167.6 DM to circulate draft agenda of AHM2005.
ongoing - SP said that five people had been chosen to attend AHM2005 booth
167.7 TD to forward Graeme Stewart's summary of Storage/Data Management
Workshop at CERN to PMB.
done
167.8 RJ to raise CDF deliverables definition with Todd Huffmann.
ongoing
167.9 TD to invite Steve Fisher to discuss R-GMA status and plans
at a forthcoming PMB meeting.
done
167.10 DB to chase up all unfilled GridPP posts at this point.
ongoing
Standing Items
==============
SI-1 ALL: to provide any news items and meeting dates for the web page to SP.
SP said she would post the press release from PPARC on Service Challenge 2
on the news page. (This was done on Tuesday).
SP asked if RM could provide her with a quote for the press release
on EGEE conference in Athens. (NEW ACTION)
NEW ACTION 169.1: RM to provide a quote for the press release on EGEE
conference in Athens.
SI-2 Production Manager's weekly report of issues.
1) SC3 preparation is happening slowly at some of the Tier-2s.
2) LCG 2.4.0 upgrades have seen much quicker responses from sites.
But still not all sites are going to succeed in the specified period.
The PMB proposed that Friday 29th April was defined as the deadline for
all UK sites to upgrade, thus defining the "LCG 2.4.0 upgrade metric".
This followed JC's e-mail to the testbed on Sunday noting that there were
only another 3-4 days left to upgrade to 2.4. Subsequent e-mails confirmed
this was acceptable.
A new metric is planned relating to installation of the VO-accepted
software at each site. The preliminary suggestion is that the
"VO latest software installed" metric is defined as Friday 13th
(subject to discussion at the next PMB meeting).
3) EGEE-3 presentations provided more information on the gLite
transition. There appear to be several competing models.
Change should only take place if gLite is better than LCG2.
TD noted that this was understood, but "better than" required detailed
definition component by component.
AOB
===
TD noted that the UK would not provide funding for LCG Phase 2 after all.
This was stated by PPARC at the RRB. TC noted that a number of other
countries would not be contributing: this is being addressed at CERN.
SP has heard back from JPhysG and two members of their Editorial Board are
reviewing the papers and will return comments shortly. (This was done on
Tuesday).
TD said he was having some difficulty arranging a suitable time for
GridPP Collab. Mtg. in Glasgow in 2007 and would check IoP dates before
deciding a definite date.
TD & SL will review minutes from face-to-face meeting.
ACTIONS AT 25TH APRIL 2005
==========================
157.2: RJ - review experiment computing models for Tier-1 to Tier-2
interactions in the UK.
157.9: RM to report on direction of GridPP in terms of middleware.
157.12: SP to provide standard template talk for schools.
159.11: RJ to make Andrew fully aware of ATLAS plans
159.13: TC to provide LCG input for GridPP13 talks
159.14: RM to extract Middleware issues from GridPP12 talks
159.19: DB to make a list of all GridPP people and their roles
167.1: JG to report back from networking meeting on service
challenge activity/organisation.
167.2: JC, with input from Stephen Burke to maintain a list of middleware
issues that should be addressed
167.3: RJ to define UK contacts for each experiment who will
respond on experiment operational issues.
167.4 TD to discuss with RM revised funding policy for GridPP people
attending HEP SysMan meetings
167.5 DK to define a possible technical documentation role.
167.6 DM to circulate draft agenda of AHM2005.
167.8 RJ to raise CDF deliverables definition with Todd Huffmann.
167.10 DB to chase up all unfilled GridPP posts at this point.
168.2: JC & AS to discuss and finalise production grid milestones &
metrics for Tier-1
169.1: SP said asked if RM could provide her with a quote for the press
release on EGEE conference in Athens.
STANDING ITEMS
==============
SI-1 ALL: to provide any news items and meeting dates for the web page to SP.
SI-2 Production Manager's weekly report of issues.
|