Meeting Notes for 23 April 2008 Status of i2u2 grant Release of Refactored Code Bugzilla e-Lab usage statistics The End

Meeting Notes for 23 April 2008

Present: Bob, Tom L., Tom J., Mihael, Liz, Eric, Marge;
Absent: Dan K, Mike W., Randy, Kris, Joao

Status of i2u2 grant

Marge gave a report on her conversation with Kathy at NSF. We have a three year grant! We will be directed to concentrate on the four existing e-Labs -- Cosmic Rays, CMS, LIGO and ATLAS -- rather than adding any new ones. We will need to do a lot to understand and demonstrate the educational value of the project. Tom L. and Jean Young can work together on evaluation. We have support for 1.4 FTEs for IT.

Everybody said "huzzah!" Really, that's what they said. And then we discussed what "huzzah" means.

Tom L. believes that the study of Detectable Forbush Decrease Searches (sun bursts affecting cosmic ray flux) mentioned in the April 9th minutes can bring together a community of students/teachers with a common goal and provide a good way to get data for evaluation.

Release of Refactored Code

Mihael deployed the refactored code as production over the past weekend. We had to iron out a few minor glitches, but it has been running well and we got our first bug report using the form Eric and Bob have worked on. Before deployment, Bob confirmed the last two bug fixes related to geometry. Mihael will subscribe to help@i2u2.org so he can get the bug reports from the form (and/or Eric will add Mihael's address to the form). Tom J. will write something in the wiki about how to subscribe to the help list, in the ForNewMembers page.

We will discuss next week the issue of creating redundancy using resources at other sites, including the University of Florida. Mihael was able to set up a backup site a UChicago so that Bob can give his workshop on April 26th when the Argonne servers will be unavailable.

Liz wondered if we should be trying to add code to CMS that at least describes the studies and has a link to OGRE. This would be like we had before Tibi made the version that integrated with the old code. Eric also suggested that we add code that will display a warning if the code the user is accessing not yet meant for production. There are various ways to trigger such a message, based on host name, URL, or in a configuration setting. Liz will put that item into Bugzilla so that Mihael can work on it. Eric will also add this to Bluestone.

Bugzilla

Mihael wants people to put any bugs into Bugzilla rather than discussing them in e-mail. Or rather, he does not mind discussing them via the e-mail list or in the telecon, if that helps, but they are not formally "action items" if they have not been entered into Bugzilla. He decides what to work on based on the list of tickets in Bugzilla.

There is a link on the front page of the CI wiki to Bugzilla. Please use it to enter all action items dealing with the code for the e-Lab or analysis tools. Instructions for getting an account on Bugzilla are at ForNewMembers#Bugzilla

It is possible to create URLs to the Bugzilla site to create various lists of tickets. Eric suggests adding a few such links to the CI developers wiki. He might even create those links....

e-Lab usage statistics

Bob would like some e-Lab usage statistics. Mihael suggested Google Analytics. Eric suggested Webalizer, which he's already familiar with. Someone recalled a warning from Tibi or MikeW that ANL may not want Google Analytics on machines they manage, but we are not sure that it's a strict policy. Since we are a research project we should be able to research whether or not it meets our needs, if not on the ANL cluster then on other machines. Webalizer will take Apache web server logs and produce nice usage reports, with bar graphs and pie charts. Eric has already used it on Pirates@Home, and he will try it out on i2u2.spy-hill.net, and then configure www13 to use it. Mihael says that Tomcat logs have a different format from Apache logs, but Eric thinks it is likely that it is easy to configure Webalizer to process Tomcat logs as well. He will investigate that further.

Mihael and Eric will report what they learn at the next meeting.

Eric mentioned that Bluestone also needs usage statistics, and he already has plans to add them. The easiest place to start is simply to count the number of "slots" created in the last day or so. Each analysis uses a separate slot directory, named after the PHP session, so this will count student sessions using Bluestone, but not the number of separate analyses run in those sessions. More detailed usage information will come when Bluestone starts logging info about each analysis when it is launched and completed (or fails).

The End

Eric and Bob will try to meet later this week to talk about Bluestone before Bob's workshop at the AAPT meeting in Minnesota.

When Liz had to leave Bob took over taking notes, but then Liz's computer kicked into screensaver mode, and Bob didn't know the password, so he could not take more notes. So we decided it was time to quit.

-- Main.LizQuigg - 23 Apr 2008
-- Main.EricMyers - 23 Apr 2008