Core Minutes 4/12/2005
Present: Joanne Bogart, Anders Borgland, Toby Burnett, David Chamont,
James Chiang, Johann Cohen-Tanugi,Pol D'Avezac,
Seth Digel, Richard Dubois, Berrie Giebels,
Navid Golpayegani, Tony Johnson, Heather Kelly, Michael Kuss, Matt Langston,
Chuck Patterson, Igor Pavlin, James Peachey, Dirk Petry,
Leon Rochester, Tracy Usher
- Data handling: (Tony)
- Data Server Richard is testing the new web front end via
a secret URL before release. Tom will be back soon to complete
work on the back end.
- System Tests A new version was released, then backed out due
to problems.
- Pipeline Dan is testing a new version using Navid's batch
submission. In about a week, if everything looks good, will make request
to CCB to release it for production use. Dan is also working on the
archiving question.
- Pipeline front-end There is a new version, designed
to address some high-priority requests from Dan. An even newer version
is close to ready; expect it around the end of the week.
- GlastRelease: (Toby)
Although v6r6 built correctly on Linux, there was a problem having to
do with GRB and ROOT on Windows. v6r6p1 addresses this partway; that tag
builds all right on Windows for users [not perfectly, since a couple
random, non-crucial applications fail to build; see
RM-22],
but not for RM. (Richard) System Tests were started automatically
for v6r6p1, a first!
- CAL news:
- CalRecon reorganization (David)
See the
CalRecon Confluence page
to get an idea where CalRecon organization has been, where it is, and where
David would like to see it go.
This will be discussed at the Calorimeter meeting on Wednesday.
(Tracy) work continues on fleshing out TDS classes. Bill has some suggestions
on what the output of energy corrector tools should be, to be mulled over
by others.
- Package managers Zach will be package manager for CAL
packages other than CalRecon. David will own
that one.
- RM, Installer on Windows: (Navid) RM is
chugging along. Queue has been modified to allow 3 concurrent jobs.
He is about to enable email notification of failures in Windows builds.
The latest version of the Windows Installer is also working (previous
version had a bug, so be sure to get the new one), though it doesn't
yet have anything equivalent to the bin directory on Linux, where all
executables go. It all worked quite smoothly when Richard tried it
out on his laptop. After installation, he did a broadcast set-up from the
Gleam package and was able to run executables with no problems.
Stil to come is some support for just installing the libraries needed
to examine ROOT trees. Toby suggested making a container package for them,
but Heather and others believe we might regret this since we could lose
consistency. The libraries should come from a particular EM or GR tag.
Navid had to move the MySQL dbs to its own partition since another
application, elogbook, was eating up the space.
In fact he didn't quite make it in
time, but was able to recover. The database is completely restored
and up. (Richard) This database is more and more a critical piece of
the infrastructure. (Navid) is already using Goddard as a back-up.
The server there grabs changes every few minutes.
- Calibration news (Joanne) will make a
new tag of CalibSvc shortly which will, by default, give up the ghost
if a requested calibration can't be found. A new job option will make
it possible to return to the old behavior, which might be appropriate
for some circumstances, but not for Pipeline runs.
A newly tagged version of rdbGUI (check out container rdbGuiRelease
v1r7 or use it from SLAC Linux) is available. It makes it easier to
select the correct DB schema and has a new "copy row" function.
Discussions are under way to decide upon an implementation of
write protection for the area where calibration files are written.
Several people have permission to write files there, but they should
not be able to edit or delete them. There will be a calibration meeting,
probably this Friday at 11, to discuss this and other issues.
- EM releases(Heather) p10 was released
late last week; expect p11 soon. It will have TKR tags for
calibration support, some CAL stuff, and small updates to RootIo.
Don't forget to tell Heather about anything that should go in: what
tag of what package, what wonderful new features it brings us,
related JIRA issues,..etc.
What exactly is the path a new package tag should follow to get into an
EM tag Leon (and others, no doubt) wondered. Used to say it should make it
all the way to a tagged GlastRelease before being promoted in EM, but at
least in some areas, like LDF and calibration, this isn't practical since the
code doesn't get exercised in GlastRelease to the extent it does in EM.
Another problem with this is that GR tags have not been that frequent.
However, there was general agreement that we probably should go to a
schedule of tagging whatever is in GR HEAD about once a week.
Heather would like an easy way to tell which release tags correspond
to releases used in production. (Richard) Maybe this is something Navid
could keep in his database.
- ACD: (Heather) is finding some spare
cycles [hard to imagine where] to spend on ACD calibration issues.
- Data-taking: (Richard)
Two towers are in the grid. There will be test runs through this week,
calibrations probably on Tuesday of next week.
- Science Tools Checkout 2: ..is done!
- Miscellaneous announcements: (Richard)
There will be a Subsystem Managers meeting Wednesday and also
a LAT All-hands meeting/celebration.
J. Bogart, Last Modified:
01-Jun-2010 15:46:57 -0700