GR status: (Richard) Two GR tags are in the hopper:
When each is ready, we need the usual diagnostic simulations runs.
(Heather) GR tag v15r51 has the new OmniOrb. r52 will have the new LDF, v07-01-01. She plans to use a version of xrootd client which has not yet been incorporated into any tagged ROOT release. Tony Johnson has tested this version via the skimmer using ROOT v5.18e. We plan to upgrade to ROOT v5.20 and will include this new version of the xrootd client. Richard reports that there are still occasional problems when the servers lose a lot of memory quickly and have to be restarted. However, this new xrootd client is an improvement.
Overlays (Heather) Work on overlays is proceeding on the r50 branch. (Leon) has been exercising the system and has run into some problems:
(Tracy) is working on the new scheme (applying calibrations to skimmed periodic triggers first before merging). There are four pieces to this: TKR, CAL, LDF and ACD. They're all working except for ACD, which is in progress. He hopes to finish it off in a couple days. Still to be addressed: merging information in case the same tile or ribbon is hit both in the MC part of the event and the overlay.
(Richard) Overlay is needed by early December in order to regenerate IRFs.
FSSC Report: (Eric W.) Made some fixes to ScienceTools v9r8p2 and made a tarball of the result. Distribution awaits completion of ongoing rigorous testing.
Diffuse response: (Richard) At the end of next week, expect to reprocess FT1 with new diffuse response columns. Jim Chiang still has a little work to do and Tom needs to write a pipeline script. May hold off on the full production run until the new diffuse response model is ready (about 3 weeks).
Disk space: (Richard) has reached agreement with Luca that so-called 10% solution will involve deleting recon files older than a month.
Our next purchase, originally scheduled for January, may be moved up to take advantage of the excellent terms negotiated by Atlas ($75/Tbyte), which expire Dec. 12.
Documentation: (Chuck) sends this summary:
This past week was spent working on data access "help" files. So far, I've posted pages for the following:
- Data Catalog
- Data Catalog Line Mode Client
- Merit Skimmer
- Fits Skimmer
- Astro Server
I'm currently working on one for the Download Manager, and will also be meeting with Seth this afternoon to get input, so we can expand the introductory material for the Astro Server's primary target audience.
Still to come are the pages for the WIRED Event Display, and History, though I have done some preliminary work on both. Once all of these have been posted, I will then be able to make another pass to flesh them out and tie them together as appropriate. So this section of the workbook will continue to evolve in the weeks to come.
Yesterday, Andreas and I completed a preliminary page detailing the setup for running iLat from the desktop. (See iLat Desktop Installation (preliminary)
Jim Chiang also updated his gtmktime file to include a new, roicut parameter and, as usual, there have also been a number of miscellaneous corrections and updates.
I also plan to make a new tag for the WB later today, so Berrie can update the mirror site.
SCons meeting (Heather) We had a meeting last week dedicated to SCons and much information was transferred (see minutes), so we've instituted a standing meeting at this time (Wednesdays at 9 AM). Anyone is of course welcome to attend.
Navid report (summarized in post-meeting submission below; thanks, Navid!)
GR and Externals (Emmanuel) All packages build one-by-one. [ Congratulations! ed.] Next step is to build them all with one command, which he is planning to do with GoGui. He has the RHEL3 executable installed on a local machine but needs to install some other tools. (Joanne) Be sure to pick up the latest version of GoGui, 0.8.6; explanation below.
obf external library still needs work. (Navid) will take a look at Bryson's handling of the issue for the similarly-structured fsw external in CHS-scons.
GoGui (Joanne) discovered that the immediate crash Heather encountered when she tried to use the Windows build of GoGui had nothing to do with Windows, really; it had to do with starting fresh, with no previously-saved configuration from an earlier GoGui session. The bug is fixed in GoGui 0.8.6. Builds for RHEL3, RHEL4 and Windows can be obtained via ftp.
She's in the midst of adding an exclude-package function to make use of the new --exclude option. Most of the bookkeeping (there is quite a bit to keep track of what happens when a package is added or deleted, when there are several versions of the same package, etc.) and visualization groundwork are done; she has yet to write the code that invokes SCons with excluded packages.
previous | minutes index | next |