Agenda and Minutes, 2 May 2007
Agenda: Wed, 2 May 2007
Minutes
Attendance: Marie Anne Bizouard Stuart Anderson Andrea Vicere Livio Kipp Cannon Brian Moe Adam Miller Nick Fotopoulos Patrick Brady Duncan Brown Xavi Siemens John Zweizig Kent Blackburn David Meyers Ed Maros Steve Fairhurst Greg Mendell Filippo Grimaldi John Whelan Keith Thorne Warren Anderson Data transfer: Stuart reported: No data has transferred yet. There were some communication issues. There has been no BBFTP installed at Caltech yet. Andrea wondered if all the appropriate information has been received by all of the appropriate people. He was worried that messages which were sent may not have made it through. Andrea will circulate this information via a web site. Andrea thinks Livio will be ready to transfer from Caltech to Cascina. In long term, will set one up from Lyon to Caltech eventually. What about the reduced data set. Benoit was working on generating this, but ran into some disks problems. Will be transfering data from WSR8 run as prototype. Stuart summarized that this will be a set of perl scripts pushing from via an ssh authenticated and then ftp. Benoit will try to make a prototype frame with the type zero DQ information inside the frame; there will be an ascii segment file with the DQ information. Benoit will provide a simple frameL program to pull that out. Consider Stuart to be the project manager for the data transfer from Virgo to Caltech. Next steps: Andrea to post on the web information about bbftp so that things can actually move forward. Expect to have this working by next week. Andrea said we need it. The publication into LDR at CIT will be present; diskcacheAPI would find data at Caltech. Doing queries against segments would wait until Ben comes back. With Benoit's sample code, Igor and Duncan may be able to get this into the database too. Marie-Anne: asked to know which information will be extracted from the frame and which will be provided later. Benoit will explain this on the web page. Announcements: LSCSOFT: problems with user env scripts ...... need to be sorted out today. DMT: John had nothing to report on DMT. Filippo has circulated the mail about the build of DOL for FC4; rpm in there including src, spec, binary, etc. Ed sent a link to a new version of framecpp Adam reported that there are builds of dol, root, framecpp, user-env. Has installed on marlin and done some basic tests. Adam will send the new spec file to Ed Maros and he will encorporate the changes into the spec file in CVS. These changes will then be available in the future releases of framecpp. Glue: One of the things coming down the pipeline from glue. With the next version of condor will be able to have dags self limit different types of jobs. Just need to know the appropriate knobs. Stuart thinks we just point at a different config file. Breadth versus depth can be obtained by getting a new dagman binary. Duncan: Maybe also set priorities at different levels of the dag, but need all jobs in the queue for that to work. Stuart asked about the ranking. Priority within a single users jobs. Rank may put different users in different ways. LDAS: Working out the last big fixes before the release. LDR: Kevin started transferring data that Igor wanted. Admintools. Nothing to release yet. When will we have something? Any update on the small file, gridftp support? LAL/LALApps: Drew is working on getting the gzip compression into the xml reading and writing tools inside of liblalsupport. There will be a required argument to the reading and writing and is almost ready to go now. Pylal is already set up .... LDG: Greg will have a fix for pyGlobus on Intel Macs later today. Adam and Nick have provided a fix and this looks like it will work. Keith also has a problem with pacman. Greg has a fix for the pyGlobus issues and will circulate them. Keith will try to understand the issue he is seeing and forward to Greg. MatApps: Keith reported that some stochastic people want to use mkframe and metadata. Most of the work lately by Ben and Justin to allow people to access data using the NDS system into a matlab client. Onasys: Nothing to report. When glue is released, will put out a release of Onasys as well. OSG: David has been documenting configuration changes for large scale workflows on OSG 0.6.0. He expects to have a document that can be used by sysadmins on production class machines. Will try to do a production level analysis ready to run. There is good support from the Globus community. David would like to run at UWM with web services. AOB: Duncan will be chairing a committee on future generation operating systems and that sub-committee will report back in a few weeks.
$Id: 070502.html,v 1.2 2007/05/09 18:22:27 patrick Exp $