Date and Location -------------------------

  1. July 2005
  2. :00 - 17:30

Attendees ---------------------------------

Di Qing






Ron, Mark




John Gordon, Andrew Sansum, Martin, Derek

Jamie Shiers, James Casey, Gavin Mc Cance, Patrica Mendez Lorenzo, Sophie Lemaitre

Michael Ernst

Lassi Tuura, Nick Brook

Subject -------------------------------------

__________________________________________________________________________ Jamie: Just to clarify that currently only Oracle is supported as a backend for FTS. MySQL is planned for next release (October?)

Jamie: Workshops - not much feedback on workshops, it looks like it's unlikely for anything to happen in September. There's no room at CERN which can hold more than 10 people. If we did do it outide CERN, we need good numbers by the end of July - currently only three response, so not enough.

Jamie: Following information provided by sites, this is good. We should present a status table every PEB showing progress at the sites. We need target dates for component installation.

Lassi: Can we have a status report on mailing lists/wiki every day/few days.

James: I've been starting one

Site Status:

ASCC: Still trying to get castor srm working.

BNL: Need to update to new version of FTS next week.

FNAL: We're working.

GRIDKA: They're ready. Small issue with creating directories.

IN2P3: Transfers working. No more than 60MB/s please!

INFN: Luca - no idea of status. Problems with cooling @ T1 - had to turn off all machines, need to decide on what to do to restart them.

NGDF: Anders- problems installing DPM - Ready to go. Question on workload.

James: Load generators are on the website - if you want to run them, you can, if not we can. I'll add people's DNs to the list of managers so you can shut down/start up your own channels.

PIC: Testing with CMS load using gridftp. Not much load via FTS on SRM now - would like to see some testing of it.

RAL: Tier1 ready now dCache running - finishing up the network testing. FTS installed - Tier1 transfers to edinburgh done. Want to understand what the interaction between FTS and Phedex transfers�

SARA: dCache working, not getting the full rate we would like - currently 120/130MB

TRIUMF: dCache working - we see 4 simul. transfers . FTS installed, LFC not installed. Working on tape integration for dCache

James: Yes- this was a FTS issue, things were timing out, so I reduced the number of sim. transfers

Lassi: From the CMS side of things, we see enormous number of transfers failures. See high rate of failures. See about 160MB/s throughput max, but the rate is higher� Transfers to gridka don't see to be working very well.

Michael: We see the same to DESY.

James: We see about 10MB/s on individual transfers to DESY

Michael: And we do 6 at time, which correlates.

James: We want to move to castor2 - we have new pool nodes, and are laying down a new data set. Once that is done, I'll start some transfers. To move people over is DN by DN. When people want to start transfers from the castor2 pools, I'll add their DNs.

Jamie: We need to agree on what times we do the transfers as the throughput phase, and when the services work. We need to do the setup phase now under stable conditions, and not step on each others toes.

[discussion on how to best achieve the goals of the service challenge.]

Andrew: From RALs perspective, we'd like to start with the FTS, test the underlying technology. Then once we've achieved the goals, we add in CMS transfers over the top.

Jamie: Can we agree to start on Monday�

Mark vdS: Do we wait until Monday

Jamie: We use Monday as the day from which we start counting.

Ron: that's ok.

Lassi: From CMS point of view, sites can decide when to do transfers.

Now to talk about Service phase:

Jamie: Need to try and define as closely as possible the service phase.

Mark: Still unclear what is needed for the service phase. What services should run on the storage and compute element parts.

Jamie: Tried to produce the first schedule. We need to iterate, and fill it in.

Nick: Should I talk 1-1 to you to see about the schedule.

Jamie: Best off-line since ALICE and ATLAS are not represented.

James: You can contact people at CERN 164111, or at <> during the service challenge phase.

-- JamesCasey - 06 Jul 2005

Edit | Attach | Watch | Print version | History: r3 < r2 < r1 | Backlinks | Raw View | WYSIWYG | More topic actions
Topic revision: r3 - 2007-02-19 - AlbertoAimar
    • Cern Search Icon Cern Search
    • TWiki Search Icon TWiki Search
    • Google Search Icon Google Search

    LCG All webs login

This site is powered by the TWiki collaboration platform Powered by PerlCopyright & 2008-2021 by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
or Ideas, requests, problems regarding TWiki? use Discourse or Send feedback