Agenda and Minutes, 11 October 2006
Agenda: Wed, 11 October 2006
Minutes
Attendance: Patrick Brady (minutes) Kipp Cannon Jolien Creighton John Zweizig Junwei Cao Keith Thorne Duncan Brown Shourov Chatterji Stuart Anderson Ben Johnson Kent Blackburn Nick Fotopoulos Warren Anderson Ed Maros Igor Yakushin Michael Samidi David Meyers Announcements; Caltech cluster is down today for maintenance work today. Discussion of possible way to track the status of various clusters. Patrick said he'd like to have some way to advertize and discover the status of the LIGO data grid in a central location. Ben advertized the monitoring tool they use to keep track of RDS, etc. Ben and Brian will be in touch. DMT: Nothing to report. Glue: Duncan and Igor have put together a plan for the DQ flags, etc. Igor will have to implement the solution. Shourov noticed problems with the daily dumps. Igor is working to track down the problem. The only problem is with the first injection segment. This is not seen with a segFind query. Duncan was wondering if the online version numbers would be updated. John would prefer not to. Igor will update the server and the clients. Duncan will patch the existing segments so that they work with the new server, clients. Duncan mentioned that there is a problem with the lfn table in ligolw-add; will roll a new release. LAL/LALApps/Pylal: Nothing to report. LDAS: Still working on the port. LDR: Nothing to report. Status of data at Nemo LIGOTools: Nothing to report. MatApps: Not much to report. Shourov has got a lot of his stuff updated in MatApps. This is good. Switch 2006b soon. Keith is getting the compilation instructions. Onasys: Nothing to report. LDG Client/Server: Nothing to report. Junwei is following up on Steffen's e-mail. LIGO-OSG working group: David is working on the LIGO-OSG working group web pages under DASWG. Reported on the amount of available disk space at the various OSG sites. They are working on partitioning sites. They hope that they will be able to fit on OSG production sites in the very near future. They have work from ISI that they will try to integrate into workflow planner. They have run a full hipe on UCSD and PSU. There is no information about the relative performance. The total aggregate is about 2000 slots, i.e. about 2000 CPUs. There are many issues that the grid hides, but really do effect the throughput. Kent suggested that partioning of data into smaller blocks to allow running on other OSG sites is one of the things that we should watch over the next few weeks. Other issues: Eirini sent an update to say she is in the process of validating the calibration frames and will attend the call next week with an update. Patrick asked about LSCSOFT. Stuart reported that things seem to be working now, so they will be ready to go once the CIT comes back up. Patrick mentioned an issue he has seen with numpy and matplotlib. People should be aware of this since there may be some instabilities over the next while.
$Id: 061011.html,v 1.2 2006/10/11 18:47:12 patrick Exp $