Core Minutes 9/14/2004
Present: Joanne Bogart, Toby Burnett, James Chiang, Johann
Cohen-Tanugi, Seth Digel, Richard Dubois, Dan Flath, Warren Focke, Riccardo
Giannitrapani, Berrie Giebels, Navid Golpayegani, Traudl Hansl-Kozanecka, Heather
Kelly, Michael Kuss, Chuck
Patterson, James Peachey, Dirk Petry, Leon Rochester, Alex Schlessinger, Tom
Stephens, Tracy Usher
-
Schedules: (Richard) See this
zeroth draft DC2 schedule. Note we need to be ready to generate datasets
for background rejection by end of January, not that far away.
This is to be balanced with flight integration work. Calorimeter is
expected by late October. Current date for tracker is also late October, but
not as firm.
-
Collaboration meeting: (Richard) It's at
the end of this month, Sept. 27-30. Core software people will mostly be here,
but the
agenda is already too full to add a developers' session.
-
Sim/recon Core: (Toby) He expects to
tag GlastRelease shortly. There are several packages which should
perhaps be promoted. Those who know about their status should inform
Toby or just promote if appropriate. (Richard) Cal packages are not ready to
be promoted.
-
Science Tools Core: (James)
- Having heard no objections, will go with
naming proposals described last week.
- Richard's draft DC2 schedule is consistent with the Goddard schedule to
exercise Science Tools. Will use Observation Simulator input for the Science
Tools testing.
- Code installer: (Navid) RM has been
modified to (additionally) output a Windows list of dependencies, one of the
steps needed for the Windows installer. (Richard) What is Linux installer
status? (Navid) The hard part -- generating a shell script allowing users to
run executables without CMT -- is done. Still need to tar up all the binaries.
Will need to write something analogous to this script for Windows. Time scale
for completing all this work is a few weeks. (Richard) It may be time to get
comments on user interface end of the script from, e.g., Toby, James, Jim.
- Corba crash resolved: Marco's new
build of OmniOrb has done the trick. It is now installed at SLAC and is
also available from the
SLAC ftp
site.
∫ Flight(w, ..) dw
- EM tag status:
(Heather) She is in the process of making end-to-end tests with the new job
options files. If this goes well she'll make a new EM tag.
- Trigger bits: (Joanne) Xin's original
request has not yet been resolved: how to sensibly initialize trigger word in
the Event Header so that TriggerAlg can decide whether or not to write to the
word. We also need to decide whether there are any circumstances under
which GEM information gets written to this word (instead of or in addition to
the 7 TriggerAlg bits). To be determined offline by concerned parties.
- ToT: (Leon) Not yet ready to roll; no news at the moment.
He will inquire further among those doing the rolling.
- CAL: (Richard) In the process of
converting to use new Response Service. CalRecon is done. Richard is
finishing up dealing with sticky C++ issues in CalDigi. (Joanne) 4 new
calibration types are needed. Work is in progress; should be complete around
the end of the month.
- Em2: (Richard) It's now expected
at SLAC in a week to 10 days. It will be at least 2 weeks before data taking.
Are we ready to handle ACD data from the mini-ACD? (Heather) Yes! But it
would be nice to see some test files for ACD.
- Pipeline: (Dan) Three small projects are in progress based on current
(Perl script web interface) version. The first is to configure a database
connection pooler so that pipeline processes and the scheduler can share
connections. This will solve a problem that currently limits the number of
runs that can be simultaneously processed, since each process now requires up
to two DB connections, exhausting the allocation of ~180 rather quickly. (This
will be corrected in a future release after moving to Navid's LSF wrapper
module from Dan's simple implementation.) Alex is investigating using a free
3rd party tool for this (SQLRelay:
http://sqlrelay.sourceforge.net/); otherwise it can be done via the Perl
DBI module and Apache. This should be working by week's end. The other
projects are 1) to integrate an email-sending tool Navid has provided to send
emails to the owner of the task and to a mailing list on failed runs, and 2)
to integrate Navid's archiving tool. These require slight retooling of the DB
management module and will be completed Wednesday or Thursday. Meanwhile Matt
has been working on a Java front end which will have more capabilities than
the interface in use now. It may be ready by the end of the week.
[thanks to Dan for the above paragraph. ed.]
AOB
- Release Manager code review: Part 1
(overview) has been scheduled for tomorrow, Wednesday, at 8 AM PDT. Part 2
(close look at Perl implementation) will occur some time next week, TBD.
- Precession utility: (Dirk) We need the
ability to precess object positions from astronomical catalogs from epoch A to
epoch B, A and B being free parameters. We'll either need to incorporate the facility into
the astro package or access it via an external library.
[Post-meeting, a solution has been found. WCSlib, whose source has
already been imported into astro, has all the
required functionality.]
J. Bogart Last Modified:
01-Jun-2010 15:47:33 -0700