Meeting Notes for 21 November 2007

Present: Liz, , Mihael, TomJ, Bob, Eric, Mike
Absent: TomL,

Standing up the New Data Server

Mihael announced Rick in Systems said that he will have it up by next Wednesday. This will be data2 and data from data0 will migrate to this. Will the new machine be named data0?

Discussion of Machine Deployment

We had a long discussion of what is the best way to use the various machines. Mihael will document how he would like use the various machines in our cluster. Mike brought up the issue of whether we should be purchasing new machines to replace the old ones from Fermilab.

Some of the comments included: Eric wondered if the new machine would be named data0. Mihael does not think so. We could consider making data0 the primary www machine and call it www18. Make www18 the primary webserver when you cut over to the new code base. Convert all the other machines to run analyses and get rid of the load-balancing. Should we look around for some money to buy another server? Mike thinks he can pay for it. Should we move load balancing to the new www18? The load balancing is on the MCS cluster. It is stable and separated from the cluster.

Action Item: Mihael is going to write up his proposal for the use of the machines.

Presenting a common URL (one namespace)

Eric would like to see displayed rather than www14, etc. This would help with bookmarking. Mike suggested asking systems to do that. There are complications we need to consider. We don't want them to see www1N in the URL. Eric discussed having reverse proxies. Mihael wants to use the Argonne machine as the stable machine to forward to our main server.

Action item: Mihael will write up the requirements.

Eric wanted us to keep an open mind about getting rid of the elab directory so users could log onto Liz and Mihael did not agree. We need the two directories for elabs and ilabs. (LIz)

Target date to roll SWIFT onto production

There was a long discussion of the steps we need for this. Action Item: Mihael will list these in the Wiki with the hope of providing a refactored version for users to work with by Dec. 7th and making a final switch in January.

Bob is going to work hard to get the testing done. Tom L. had his students test it. Miissing links. Overload conditions.

We will provide logins to the individual e-Labs rather than having a central login.

We will add a news item for user on what will changed for the end user.


Mike Wilde said that SC2007 was very successful for them. They have some people who are seriously interested in SWIFT.

Eric got good feedback from the workshop.

Tom L. missed the telecon but adds this report: ELabs was involved in four different presentation fora at SC07. First, Tom L with Pat Mooney presented the CRiL, along with a poster including information on e-Labs, to an Education Program poster session-reception for some 200 participants on Saturday evening, 11/10/07. Next, on Tuesday, 11/13/07, a 1.5 hour workshop was presented in a parallel session to K-12 educator participants as part of the SC 07 Education Program. Tom L presented, joined by Tom J, Eric, Mihael and Pat M during the hands-on portion. There were 9 participants in that workshop. On Thursday, 11/15, Tom L presented an e-Lab overview in a 30 minute session in the Education Program Booth in the SC main lobby area. Many people passed by, but only 3 (other than ELabs folks) viewed the entire presentation. Finally, Cosmic, LIGO and CMS e-Labs were featured, together with a QuarkNet cosmic ray detector (not the CRiL), in a display area as part of the Fermilab exhibit on the main exhibitors floor. Mihael, Eric, Dan K, Tom J and Tom L rotated staffing the booth for the full 26 hours the exhibit floor was opened. We had a light but steady stream of traffic, perhaps 3 or 4 groups per hour, throughout the exhibit. I estimate that we spoke with perhaps 150 people.

Also: the chair of next year's SC Education Program invited us to be more active participants in planning and executing next year's SC 07 Education Program. We should discuss this. --Tom L


Eric needs to rig up code to submit to Mihael's form. Handing things off and back. Eric will be scaling back his effort in the coming months. He is taking on another job doing some consulting for a couple of months. Once grid execution is working, there are lots of changes to the LIGO analysis that he wants to fix. Needs labeling; wants to reproduce a plot that Dale's students did of 15 different seisometers did.

Dale Ingram's e-Lab is ready for use by students. He can probably stabilze it enough so the Mihael can finish the conversion to the refactored code.

CMS and using SSH instead of Globus

Mihael playing with SSH to allow running without Globus. This way we can avoid installing Globus on all machines. This will help also with CMS. Dan will keep the analysis code on his machine.

New Hire

Mike W. is going to write a job description for the new hire and send it to the list and to Marge and Randy. He suggested that maybe Mihael could the technical manager at 30%time.

-- Main.LizQuigg - 21 Nov 2007
Topic revision: r7 - 2019-05-22, AdminUser
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding Foswiki? Send feedback