Minutes for the terameeting of September 17th 2007

Present: Yannick, Emmanuel, Chiara, Mireille, Henry, Gregory, Fred (partly).

Emmanuel's summary document

CFHTLS operations
 the missing 100 catalogues mentioned by Henry are due to manipulation problems with Condor. There should be in the end 421 Wide chi2 catalogues + 8*5 chi2 catalogues. It takes about 24 hours for computing 280 chi2 catalogues. The new catalogues seem to have the correct error estimates. The first thing to do is to find the remaining missing images, and combine the catalogues to make multiband catalogues. Everything should be fixed by the end of the week. Update: Fred is rerunning the whole thing on the "official data".

MEGACAM PI
 The Soucail data seem to have been downloaded by Genevieve.
 Emmanuel will investigate again the 2MASS problem this evening. Mireille mentions that the problem occurs also in i and z. Update: Emmanuel found out that the fact that 05AF14 could be calibrated with 2MASS was coincidence; the field is too crowded and reddened to work properly with an infrared catalogue.

Developpements T05
 The commit/log thing should be done by the end of the week. The first ingestion tests will be made on Mix9. Gregory is enthousiastic about Django for interfacing Python.
 Emmanuel will take advantage of his visit to DES to ask about spherical coordinate usage in DES databases.

Software
 Chiara will try to investigate what is missing in the SCAMP XML output.

System/hardware
 efigix (and may be mix6 according to Henry) cannot send e-mails from the condor account. Fred thinks this should not have worked. Could it be possible to masquerade the "From:" field to make the IAP pass the mail? Update: the problem was caused by a different setup on efigix (by Anthony?); it has now been reverted to the default one. Note that Condor e-mails are filtered out by the IAP spam removal system (which explains why Emmanuel and Chiara had no problem receiving e-mails from the condor account).
 Fred will reorganise the power supply of the cosmix cluster on September Wednesday 17th.

Next terameeting: no date set yet