Meeting Notes for 23 January 2008

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

LIGO wiki incident

Eric reported on an incident in early January with the glossary wiki which is served by reverse proxy thru tekoa and www13. He allowed anonymous edits to talk pages (only) to allow visitors to add comments about his page reporting on the new OLPC XO laptop. Someone from LIGO cybersecurity found it and started editing and became concerned. Eric saw the changes and reverted them, and closed the wiki. The person at LHO who is now responsible for tekoa, Greg Mendel, sent Eric an email about it, but Eric's outbound email broke that day, so Eric's response did not get delivered. In the absence of an explanation Greg shut off the proxy. Eric is moving his XO laptop report to another wiki.

Eric and Dale had originally decided to serve this wiki from tekoa because a school in Washington was having trouble accessing the original, at i2u2.spy-hill.net (possibly because of the .net address). But now that it's available via www13 Dale is happy and doesn't think we need it back on tekoa.

A lesson learned from this: proxying and wikis raise issues that have to be dealt on the front and back end. They involve filtering issues for the people accessing the content, and security issues for the systems that are serving the content and allowing others to edit the content. Nervousness on both sides!

Eric also noted from this that the initial anonymous edits, which actually came from LLO (Livingston), showed up in the wiki as coming from an LHO (Hanford) IP address. It was, in fact, the IP address of tekoa, since it was the proxy. This means that the logs on the actual server will show the IP address of the proxy machine, which means that there isn't a good audit trail. The logs on tekoa may show the actual address, which would allow us to find a machine if needed, but we'd have to correlate logs from two different machines. We don't usually need a detailed audit trail, but it's worth noting that this may be a concern if we use a reverse proxy and need the IP addresses of the originating connection. (Of course that IP address could be for a forward proxy for a school or school district.)

Refactored Code

Reverse Proxy Timeout Error

Timeout through the reverse proxy was causing problem for large uploads of cosmics data. Mihael fixed this by increasing the proxy timeout to 30 min.

Action Item: Mihael will add notes to the new wiki page Proxy Issues about how to change the proxy timeout in an apache config file.

Threshold Issues in SWIFT

Mihael fixed threshold issues in SWIFT. Thresholds will have to recalculated everytime.

CMS

Mihael received package from Dan K. with OGRE in JSP. He is trying to adapt it into our generic infrastructure. He is pretty close to getting it working. He needs to set up www13 to test this because it is alpha code. Liz asked if this would be on port 9080 because 8080 currently has Nick's code changes and Amelia code. 9080 on www13 is probably not allowed through the firewall. Mihael would rather update www13 port 8080 to have the refactored code. The idea is that www13 port 8080 would be a place to try things out (alpha), while www12 would be for testing code we think is probably ready for release (beta code). www12 port 9080 will pass through the firewall. www13 allows uploads from Dreamweaver.

Problem with popups showing from the references: If the user has already opened a reference pop-up, the window is already open and does not come to the front when another reference is accessed from the milestones page.

Action Item: Mihael will make this work for now. In the future we may want to implement this differently. Eric does not like popups and the dependency on JavaScript.

Status of e-Labs

Cosmic

The version on www12.i2u2.org:9080 is almost ready for movement to www18.

Action Item: Bob needs to test uploading large datasets with the reverse proxy and bless it.

We would like to put the refactored code on www13.i2u2.org, but we need to be sure that any code that Nick was working on gets into CVS. We also need to be sure that Tom J. doesn't have any outstanding code there that would get wiped out. Nick was working on versions to allow capturing the temperature and air pressure and to analyze that data.

Action Item: Tom J. or Bob needs to look at the computer Nick was using and save any new code. The improvements he worked on should be put in a branch of CVS or SVN, or else implemented in the refactored code by Mihael using the ideas, but not necessarily the code Nick wrote.

Eric noted there was an interesting discussion about "why upload data?" by e-Lab fellows in Cosmic Rays' Diner There seems to be a morale issue. Liz and Tom L. pointed out you upload data so you can do include other people's data in more powerful shower studies. Eric suggested that the Fellows might need to feel more specific involvement in that.

Eric suggested that having scientists working with teachers could help, but knows scientists are busy. He knows a Masters student in Scandinavia who is doing a distributed computing project to simulate air showers. Perhaps teachers could talk to him? (He's already said he would be willing.) Tom L. wants to have a grad student working with the Cosmic Ray detector in the schools. They have the money, but they have had problems finding someone.

CMS

The link to the Beta version on the e-Lab list page does not work because www12.i2u2.org:8080/elab/cms/ is no longer served.

Action Item: Get rid of the link to the Beta version.

Some of the references need updating for CMS. Currently the referenes are standalone files and no longer stored in the VDC. Eric is going to help Tom L. get started with SVN so that he can edit the files that need changing and put them back in SVN. Mihael, Eric and Tom L. will use the Aquarium Room in the Wiki to discuss issues related to getting started with SVN. Tom L. will need an account on the CI cluster. Eric pointed out that he made a new page for Newbies on this Wiki, For New Members. It can include advice on getting started with SVN. (Edit: it didn't have info on getting a CI login account. Now it does. -Eric)

Action Item: Tom L. will take steps to get SVN running on his computer. Discussion will take place in the on-line Aquarium Room. Mihael will point him to the correct files to be working on to fix the references. Lessons learned from the discussion can be turned into more formal documentation, such as improvements to this SVN page.

There are some incorrect links on the cms_support_docs.html page that is invoked from the reference you go to when you click "Describe the Test Beam Detector Setup" and click on "CMS Test Beam." See http://www12.i2u2.org:9080/elab/cms/library/milestones-map.jsp The link goes to cms_support_docs.html#cms_test_beam which has links to studies. The links to studies are wrong. Mihael said the files are there, but in a different directory.

Amelia

We don't know enough to report on Amelia. They have some pages on www13, but they look like they have not been altered much from using cosmics as a template. The Amelia pages use the old e-Lab code. We would like to see these converted to the reformatted version. They have no analysis up yet. They also have no references. Mihael needs the text for the various pages, and then he can make a refactored version of Amelia. Mihael thinks it is easier for him to be given the text than having the educator developers making pages or even include files.

Action Item: Liz will contact Kris Whelan to determine if the text on the pages on www13 is still valid.

STAR

Bill Lopes needs more funding; he had a setback with his code; he did not complete the standalone version. Tom L. has not been able to contact him. Lopes knows that he needs to make a web-based version. Eric’s Bluestone tool might be able to run ROOT for STAR project. Eric and Bill would need to confer to see if this would be an easy thing to do.

Action Item: Tom L. will send Eric the STAR desktop tool.

LIGO

No recent progress. Eric has been making sure things are running, but has primarily been working on another project (non-LIGO).

-- Main.LizQuigg - 23 Jan 2008
-- Main.EricMyers - 24 Jan 2008