Agenda and Minutes, 02 August 2006
Agenda: Wed, 02 August 2006
Minutes
Attendance:
Kent Blackburn
Ed Maros
Brian Moe
Keith Thorne (minutes)
John Zweizig
Junwei Cao
Keith Riles
In the absence of P Brady and S Anderson, E Maros lead the meeting
DMT (J Zweizig)
He is still working on the calibration frame file format software,
which is about 90% complete. He is making sure version IDs, etc. are correct.
He expects it to be about 1 day to get ready to make new RPMs (both online and
offline). Other minor mods are also being rolled in.
Glue - No Report
LAL/LALapps - No Report
LDAS (E Maros)
They are working on 64 bit version, which is going well. They are testing
32-bit <-> 64-bit object communication. The FrameCpp testing is done. There is
a problem with TclGlobus causing core dumps. They have finally moved to TCL 8.4.13.
J Zweizig said he had been trying to upgrade to TCL 8.4.13 under Solaris but with
little success. They will send him information on command-line arguments, etc. for
the installation.
LDR (B Moe)
There have been snags in testing the new version and they are not comfortable
releasing it. What were thought to be simple problem have proven to be more subtle.
Next Monday is still a possible release date, but perhaps not probable. The changes
include modularization and metadata implementation.
Ligotools
P Shawhan has left for his new position in Maryland. J Zweizig reported that B
Johnson added a firewall at LHO that has hampered some Ligotools programs as he
closed all outgoing connections. E Maros said that the new Ligotools version
P Shawhan was working on had (mostly) persistent sockets, which would solve
the problem. If need be, it was suggested that the firewall be opened to allow
outgoing connections to a restricted IP list.
K Thorne reported that he had prepared a web page detailing how to make
development compilations of FrContrib software.
Matapps (K Thorne)
Not much to report. Waiting for new UWM cluster to come online to test
MATLAB setups.
Onasys - No Report
LDG (J Cao)
K Blackburn asked he had made progress on the certificate scripts, web pages.
The existing instructions say to use iVDGL (which is obsolete) and needs to be
changed to OSG. J Cao had not done this yet, and was urged to get things updated.
J Cao said that P Brady had said he would look into it, but he had not heard
from Patrick.
K Thorne noted installation failures for LDG 4.0 Client Professional on RHEL-3.
The Professional version is needed for pyGlobus (LSCsegFind, etc.) J Cao said the
problem is that the RLS Python client only exists for Python 2.3, but RHEL-3
only has Python version 2.2. He was looking into a solution.
Other Issues
K Riles provided a heads-up on the need for support of version numbers for DQ Flags
in LSCsegFind, etc. He would like to see LSCsegFind upgrades to all selection of
DQ Flags for version # (or by 'latest'). I Yakushin will be working on this when
he gets back from Europe. However, he may be swamped with other things. K Riles
warns that solving this problem may need to be a priority for Yakushin. K Riles
noted that is currently a problem with inserting older DQ segments into the database.
B Moe was asked about the status of the new UWM cluster (Nemo). He said it was
now up, but was down for a regular maintenance day at present. He said the grid logins
for the new cluster will remain hydra.phys.uwm.edu. The original hydra cluster is
now offline.
K Thorne reported that presently jobs submitted to the CIT condor over Globus
were not working (J Cao reported same problem). The CIT people have not found any
such problems in their testing of condor submission over Globus at CalTech.
$Id: 060802.html,v 1.2 2006/08/03 16:57:43 kathorne Exp $