Teraminutes of the meeting of Friday 19th May 2006 present: Fred, Yannick, Henry, Emmanuel, Chiara, Laurent, Anis, Anthony, Mathias

Wircam counts Chiara, Henry

 discussion of counts Wircam and what is the origin of the excess in the counts. Probably the excess comes from spurious objects around the bright stars and the some masking will need to be done [postscript: it seems that masking is not enough and that there are fluctuations in the sky background which cause a lot spurious detections].

 is the magnitude limit approximately what we expect it should be be? It is it compatible? (for Henry)

 the counts from Lihwai Lin are also above the counts from Angela

 Yannick ask Henry and Chiara to provide aslo the difference between Angela and Sanders and Angela and LihWai, not anly the galaxy count themselves that do not clearly show the amplitude of the discrepancy.

 the origin of the excess in the counts probably comes from the spurious detections around the star trails; these trails should be masked.

 Henry will send an email to Loic asking once again that he provides us with the sky-subtracted data

Retina problems for MEGACAM imagesChiara, Emmanuel
 The problem has been identified: the new retina is too efficient and triggers on pieces of objects, especially on good seeing images. A new training has been started and will be tested during the next week. Deadline for release: end of May.
 because of the higher efficiency of these new retinae INTERP_TYPE should be set to NONE (as for WIRCAM).

Other matters : All

 Fred has started to look at the problems of automount and will install a test autoutmounter on pix1 which is not used Spica or by Wircam.

The forum : All

 The forum is primarily for communications with people outside TERAPIX, i.e., not for communications within TERAPIX

 the forum moderators are; Laurent (pipeline); Emmanuel (his software); Chiara (Wircam, QualityFITS); Henry (catalogues and releases) yannick (survey operations), Fred (hardware/cluster, QualityFITS)

 For .dotproject, Emmanuel and Yannick will monitor all the tasks submitted to Terapix. Yannick ask everyone to use dotproject for ALL and ANY kind of actions relevant for Terapix activities.

 Deadline has to be set for the installation of dotproject/forum; should be public the first week of june

scamp: Emmanuel

 four bugs have been found in the pattern-matching code which can affect the behaviour in very crowded fields

status Megacam data: Laurent

 1270 images have been processed; 1/3 of the very wide have been processed. Yannick isnpected all of them (Qfits-out and SPICA/Scamp output pages) and wrote notes for about 50-60% of them. Notes are written only for those having a problem (minor or serious)

 the QF L02/L05 have been re-started wednesday

 the Lihwai Lin QF restarted wednesday

 Tully, Annie Robin data retarded on friday 18

 postfix wasn't restarted

 reboot clix; /tmp was not cleared. Fixed during the terameeting.

 multiple ypserv forks were responsible for clix crash last week. Fred is now monitoring the activity of ypserv.

status of pipeline: Laurent, Yannick

 The VW on mix3 was not restarted by Fred on Wednesday. Fred missed the email from Laurent. This should be avoided.

 restart of the pipeline will be managed by Yannick and Henry. Once a problem is raised and Spica has be to restarted. Yannick will first do it with notification to Laurent and Henry via teralist. If no notification was sent, then Henry should do the restart and notify.

status of orders: Fred

 mix8,mix9,fcix2,efigix have been ordered by Fred who split the disks in different order forms in order to use EFIGI money and keep fund for anthony; they should arrive at the start of June. Once arrived Fred will notify teralist since the cluster will be stopped for installation and new cluster hardware configuration.

panorapix; anis

 it is now possible to display images and adjust contrast/brightness control with shaders and opengl (but not both simultaneously);

 the image display will now be implemented in OpenGL

 Emmanuel suggests the possibility of using the shaders to perform the byte-swapping

 displaying the image only using opengl in 32 bit doesn't work

Anthony's Thesis

 15,000 euro from EFIGI (transiting by TERAPIX) will be used to fund the PhD. Delphine is looking at the "IR-like" option instead of IE. If this is successful, Anthony will be paid during 6 months as IR, otherwise he will be paid as IE during 10 months.

 Fred has called option+ to tell them that the processors in mix5 may be defective but has had no response yet;

 mix6 passed 26 iterations without crashing; this machine can be considered stable

 pix7 will be reformatted once Laurent has finished cleaning the database and the disks

 normally the problems with /tmp not being cleaned at reboot has been fixed on all the machines by Fred.

next tera-meeting: no Tera-meeting next week because of the Ascension (likely 2 days off between May 25 and 26).