Core Minutes 1/11/2005
Present: Joanne Bogart, Anders Borgland, Toby Burnett,
James Chiang, Johann Cohen-Tanugi, Seth Digel, Richard Dubois,
Warren Focke, Riccardo Giannitrapani, Berrie Giebels, Navid Golpayegani,
Tony Johnson, Heather Kelly, Michael Kuss, Matt Langston,
Chuck Patterson, Leon Rochester, Tracy Usher
-
Root 4: (Heather) It's looking good!
There is one small remaining glitch involving calibRootData that should be
easy to
fix. She tested reading and writing in various combinations and compatibility
looks ok. There is a non-fatal complaint in one circumstance (reading with
new ROOT certain files written with old, Linux only), but this is not
a reason to wait. See more detail
on the
ROOT 4.02.00 Checklist page.
-
Pipeline: (Richard) Some problems
having to do with job submission to the right (no AFS access needed) queue
have been fixed. Pipeline (manager??) is now running at 1 minute intervals.
Dan will re-establish a test at each stage for output of previous stage.
Next task after that will be to get archive going. Meanwhile Matt has been
working on an update to the frontend, hopes to finish it off this afternoon.
It will make it easier to find current runs in the rather long list.
-
FRED: (Riccardo)
FRED now has a working button!
For details on all recent FRED developments (which tag of what to use,
explanation of a couple small oddities, etc.) see Riccardo's comprehensive
update page.
-
New RM: (Navid)
He continues to work on the Processor. Windows LSF batch is
working — to a
degree. It is possible to submit jobs from Linux to it; it's not
always possible to find the log file. (Richard) Jim has put together
a stripped-down Release Builder on the Terminal Server which can build
a single release upon demand, though without convenient access to
log files, etc., using MRcmt.
Recent GlastRelease LATEST builds, starting yesterday afternoon,
have lots of compile failures. This is probably due to a latest (in time)
tag for Event, v9r13p1gr4, being picked up. Many package versions in
LATEST depend on features of Event v10r*. After the meeting Navid
described the algorithm RM uses to decide which tags to include in
a LATEST build:
The way the RM picks up tags for LATEST is pretty much the same as the old
RM with a small change for branch tags.
Currently the RM looks at all tags for each package and sorts them in the
order of tagging. It then goes through that list and depending if this is
a compile for ST, GR, or EM it removes a few tags. Removal is as follows.
Any tags that don't follow vXrYpZ(gr|st|em)W rule are removed
Now if the compile is a ScienceTools compile it looks at the gr|st|em part
and removes all tags that have 'gr' or 'em'. Same goes for GR (remove
'st' and 'em')...
From this trimmed list the latest tag will be used for LATEST builds.
Discussion among interested parties will continue offline.
Several additional RM features were requested, such as that a
failed checkout be marked as a failure.
[Reminder from last time: requests for further enhancements
should take the form of JIRA issues.]
-
Data Handling/System Tests:
(Tony J.) Plots are now in. Work to add statistics comparisons is
in progress. Started to integrate with Matt's sign-in service. They
hope to have all this ready to go sometime today.
∫ Flight(w, ..) dw
-
Tower A: (Richard) It will be
shipped Friday, probably arriving Friday afternoon. First order of business
will be to complete Tracker testing. This may involve Eduardo & Co. (and
SAS software). Luca should be arriving around the same time, perhaps
over the weekend.
-
EM tag prospects: (Heather)
She would like to make a new tag soon, maybe tomorrow, to pick up various
bug fixes by Leon, perhaps a new version of CalibGenCal, and, ideally,
the new ROOT.
Coming Soon
- Data Handling Meeting at SLAC:
Will take place Thursday and Friday of this week at SLAC; see the
agenda. Roughly speaking, mornings will be used for
articulating requests; afternoons for figuring out how to implement them.
Julie, Navid and Heather will be in town.
J. Bogart Last Modified:
01-Jun-2010 15:47:31 -0700