Minutes for the Terameeting of April 1st 2005

Present: Fred, Antony, Laurent, Jean-Christophe, Yannick, Mireille, Henry.

Spica, status of release (Laurent)

 All the deep stacks generated with the back-mesh setting of 256 have to be re-done (with the large back-mesh, sky residuals are too large)

 Status of processing as of Friday, April 1st 2005: mix2 (40 stacks left); pix9 (20 still to do) image-by-image (600 still to do)

 It will take 3 weeks still to finish all the image-by-image processing

 It will take around one week to finish all the stacks still pending; then we have to restart the processing marked 'redo'.

 There is still a problem where the swarp crashes with very large numbers of images. Emmanuel is investigating with 'valgrind' the orgin of this problem and hopefully we know shortly. However, with the large number of images involved, this test takes a long time).

Quality assessment (Emmanuel, Yannick, Henry)

 Henry will do the catalogue extraction and stacking for d1 and redo the colour-colour diagrams as produced for d2.

 For the wide fields analysed there is a systematic scatter in the stellar colour-colour diagrams between the fields. Henry will try to calculate the offset between d1 each of the wide fields, and apply this offset to the catalogues.

 Note that there are two u*/wide fields for which scamp did not correctly calculate the astrometric solution

 Very wide (JC). JC has investigated carried out some preliminary quality assessement tests in the wide field data.

 Right now there are no very-wide fields with three filters. Could we implement some method to automatically cross-identify ldacs produced by spica to automatically produce the colour-colour diagrams? Fred had some suggestions.

 Note that in the very-wide survey there a lot images with very bad photometry: for the very-wide detection is always priority.

Data transfer (Jean-Christophe). CADC have only downloaded the image 'test', and have not yet started any bulk transfers. They say that they are ready to download the data, but not to archive them

 We will maintain pressure on CADC this week to start downloading and integration into the archive as soon as possible

Hardware (Fred, Yannick, Emmanuel) Fred will replace Liping's computer with setix (Laurent's old machine).

 We will also investigate if we can buy a new machine for Liping from the ERA credits or perhaps use one of the 'maxdata' iap machines (Liping uses her computer just an x-terminal).

 pix1 (Fred) Fred has changed one of the raid disks from pix1 (Western digital 100GB).

 Question for Gilles: how many of these replace disks are left? We should prepare to replace these machines (pix1-4) as soon possible, as many disks will start to fail soon (these machines are all three years old)

 mix4 (ex mix2): Yannick now reports that he was able to swarp all week on mix4 without any crashes (six large swarps sucessfully completed)

 Fred made several changes to timings on the pci-x bus but it is not clear why stability of the machine has been improved.

 the /raid1 partition on mix4 still doesn't work properly.

 mix5 (Emmanuel, Yannick). The demand for the machine has been posted on the CNRS web site, and we are waiting for offers from vendors for the machines.

 The purchase decision will be taken on Monday. At this point we should also order the ftp machine and the fibre channel array.

Florence Durret data (Yannick, Mireille) CCD-0 missing because there was a problem with the astrometric solution: we can re-do this stack using Fred's astrometric template. Mireille will investigate this.

skywatcher (Jean-Christophe)
 Web interface version of sky-watcher demonstrated by Jean-Christophe.

 Some discussion about the design of sky-watcher and the configuration file, which turns out to contain all information about all the survey fields.

Next Tera-meeting: