Present : Mireille, Laurent, Emmanuel, Frédéric, Anthony

Late : Chiara :-P

Absent : Jean-Christophe, Henry, Yannick

Problems on runID 03AH05

Images of QRunID 03AQ05, reasons of the problem must be investigated. Could be:

• errors in the reference catalog (ameliorations with 2mass) ?
• observations done during a bad night ?
• other ? Other fields of the same QRUNID (03AQ05) should be checked for the same error to help to isolate the problem.

Bug in Swarp

The bug found by Henry was known by Emmanuel. It inverts two lines. The bug does not come from threads so Emmanuel does not know where it comes from. This bug is not repeatable. It does not always happens in the same place. Emmanuel will investigate. It might be a problem with ds9 ?

Cluster

mix2 and mix4 will be inverted again so the final configuration is the following:
 mix1 and mix2: biopteron
 mix3 and mix4: quadriopteron

To do so, jobs on mix2 must be killed. Frédéric will run again Anthony's process next week. It needs several days to compute. It should be over soon so Frédéric will stop it as lately as possible.

Frédéric will soon install the new stable core 2.6.12 on mix5 to avoid crashes. Frédéric is also currently updating opterons.

terapix should be updated. A decision will be taken later to choose to keep a debian or to install a gentoo.

Purchase

A mix6 (same version as mix5 with dual core CPUs) should be bought soon. Yannick, Emmanuel, Frédéric will see Delphine on monday to decide. The computer should be a quadri Dualcore. Its integration into the cluster will be planned with Laurent.

Transfer clix to clix2

This operation is delayed. Frédéric wants clix2's system disk to be RAID1 safe. In the meantime, system services will be transfered and tested. Jean-Christophe daemons will have to be stopped (snoopix, datatransfer, fitsverify).

PI data

Those data are on mix1 and mix2. Raw data are kept on the fc* disks, and can easily be retrieved from CFHT again. Processed data are kept on mix1 and mix2's disks.

Transfert database to dbterapix

It will be done easily by Laurent.

qualityFITS parallelization

Pending

Office movements

Chiara and Frédéric are installed in their new offices. We must check what was done in Chiara office (add of a sheet of lead on the door, etc.).

Chiara will ask for any furniture she may need and Frédéric will go and buy for both of them.

Automatic notation

Laurent compared computed notes and manually attributed notes, they do not match ! Intern astrometric note is always C and chi2 rule is never used. chi2 value from sextractor is not good. Laurent should plot meta statistics to determine tresholds for each parameter and to refine defined rules. It will be discussed next week.

FITSVerify

Emmanuel fixed bugs (TFIELD). Concerning ldac files, he will try to change format to use multidimensional tables to be compatible with cfitsio library. Improved error estimates on XWIN_IMAGE and YWIN_IMAGE shall be done by Monday.

Gaia

gaia uses its own version of sextractor. Be careful if it appears in your path, before the newest terapix version of sex !