Bugzilla – Bug 2620
CAMPAIGN: GT4 WS GRAM Loose Ends
Last modified: 2005-02-07 18:59:45
You need to log in before you can comment on or make changes to this bug.
Campaign Leader: Peter Lane People: Peter Lane Projects: Globus Toolkit Development Community Expansion GRIDS Center Technologies: Globus Resource Allocation Manager (GRAM) Definition: This campaign serves as a vehicle for tracking various smaller WS GRAM related bug fixes and small improvements that need to be put in place before the GT 3.9.5 release. Benefits: Focus on fixing a number of known bugs so they do not occur in the 3.9.5 release. Should result in improved usability and reduce 3.9.5 support bugs. Deliverables: Stability & Functionality Bugs ---------------------------------------------------- • 2552 GRAM delegation listener does not deal with proxy refresh • 2576 duplicated directory seperators ("//") in JobManagerScrip... • 2363 anonymous GRAM info *Dependent on work from student to write the client and service parts User Error Protection Bugs ---------------------------------------------------- • 2560 NPE can occur if no delegation chosen when submitting • 2540 authorized job hangs if another job user not in /etc/sudoers • 2395 non-standard grid-mapfile location fails • 2590 ws-gram service hangs on mal-formed job description Usability Bugs & Enhancements ---------------------------------------------------- • 2524 Bad invalid script response error • 2559 Errors when staging do not make job fail but submission c... • 2062 Stdout/err RPs aren't using FS mapping • 2544 RFT transfer faults cleanup Resources: 5 days
Fixed #2552
This campaign was stalled a little for the past few work days. Friday I tried to pin down 2559 but couldn't reproduce it. Stu asked me to look at a memory bottleneck that was found, since that seemed more important than #2559 (and it was since we are now able to meet our scalability goals after the fix). The code being written by Jonathan wasn't ready for #2363 on Friday. Monday I had trouble getting info about where this code was and how to use it, so I worked on other stuff. I also took it upon myself to instruct Jonathan on how to organize the code better on Tuesday. I'm hoping to finish off #2363 today, but testing may be harder than expected since it requires the setting up of a database.
I will try to figure out 2559 today since the code freeze has been postponed to Tuesday. If I can't get to it or resolve it today, then the campaign will be stalled again for probably a week.
All bugs have been resolved. This campaign should be closed.
