Core Minutes 12/4/2012

• ScienceTools: (Jim) made a fix to fitsGenApps to correctly write the sign of the rocking angle to the FT2 file.

  See Science Tools Development Notes for details.

• FSSC: (Alex) FSSC is waiting for the next ST release. In other news, there is a new man onboard who will be taking over ScienceTools duties: Joseph Asercion. Welcome, Joe! Alex will continue to be the main contact for the time being, but expects to get back to predominantly data server duties around the end of the year.

• Pass8 eprocessing (Tom G.) A new round will be starting, likely as soon as next week, to reprocess L&EO data.

• New Fermi lnx servers (Tom G.) They're installed, currently in the process of being configured. They're bigger and faster than the old boxes and have dual power supplies. Some are also hooked up to a back-up generator. Next it's up to us to move services to them. ISOC will probably go first since they already know how they want to use the three assigned to them. It may take longer to sort out the rest. Tony J. is concerned about certain services being resident on the same machine.

• Pass 7/8 (Leon) A new bad strips calibration is now available. Otherwise, he's been working on Overlay and, in that context, trying out SCons on Windows with a supersede directory; see more about that below

• GR releases (Tracy) Standard datasets for GR 20-07-00 were generated over the weekend; it seems they're ok (no bad news, anyway). We aim to push the button on 20-08-00 early next week. No infrastructure changes are involved, just "regular" code updates, so we have hopes it will go smoothly. Then we'll run off standard datasets. The only caveat is system tests.

  (Heather) Thanks to Johann 20-07-00 has the updated externals.

  (Heather) will be on a plane to SLAC on Sunday and again, heading back home, on Thursday. This somewhat restricts the times she can start up a new release build.

• New ROOT and system tests (Heather) To review: the problem is that some of the ROOT scripts run much more slowly than they did with the previos ROOT version. It's not occurring during data reads; it's connected with histogramming. ROOT's own benchmarks look ok. She continues to investigate..

• Core talk:

  • SCons RM and GlastRelease (Heather) Tom S. instituted several changes, including turning off automatic rhel6 builds, increasing CVS checkout timeout from 45 minutes to 60, and staggering the start of builds on different OSes so that they'll be less likely to collide. It's helped! Recent builds have gone through ok.

    Still remaining: the zip files permission issue. This happens only for GR and that is, not coincidentally, the only container for which builds are written to u52. Apparently it has to do with the way this disk is configured, but so far we have been unable to get any action from the Computer Center. She's tempted to go with a work-around: cron job to fix permissions after the fact.

  • Developer environments (Leon, Joanne) Leon made some forward progress on Overlay work with Windows + SCons, using a supersede folder. Then, when he added a couple more packages to the supersede area, mysterious compile problems (unable to find certain headers) started cropping up. On Windows for GR the situation is especially complicated because headers are not installed. However, as of SConsFiles-01-06-00 (tagged on Oct. 29) all possible circumstances are supposed to be handled properly. We need to do some digging.

• AOB

  • (Alex) Would like to know when ROOT 5.34 will be incorporated in a ScienceTools release. ROOT 5.26 will not build successfully on Mountain Lion; 5.34 is supposed to be ok. (Heather) needs to make builds on the remaining platforms: redhat5 64-bit and snow leopard. (Jim) Let's make those builds and then see what happens in LATEST.

previous

minutes index

next

J. Bogart, Last Modified: