Core Minutes 11/10/2009

• A bittersweet announcement (Navid) will be taking a new position on Dec. 1 with the Earth Sciences group at Goddard. This is the same group he worked with on his Master's thesis; he'll be continuing on to a Ph.D. [Congratulations and alas! ed.]

• FSSC: (Eric) Did some work on python interface for Richard's upcoming tutorial presentation at Monash University in Melbourne.

• GR: (Tracy) He and Leon will be meeting with Bill and Robert in Santa Cruz on how to proceed, timeline for Pass8.

• L1: (Richard) The pipeline is crawling to a crawl. A couple things have been observed:

  • Many jobs are failing at the MySQL connect to the calibration database. [The relevant code is in rdbModel. A version incorporating retries has been available since last December but was never incorporated into an L1 release. It is in all v17 GlastRelease tags. ed.]
  • Afs buffers have different characteristics than they used to.

  It is unknown at this point whether one of these is causing the other or whether both are symptoms of some other underlying problem. Investigation with SCS is ongoing. Suspects include:

  • Nearly-full (95%) logging disk, u15. We've seen slow-downs in other circumstances when critical disks are this full.
  • Switch to using rhel4 build rather than rhel3 build.

  (Anders) Yesterday things seemed to speed up slightly; don't yet know whether this is significant. (Richard) There will be an outage on Thursday for various reasons unrelated to this problem. Warren will take the opportunity to switch the logging disk to u52. We will also be using a newer GR. (Anders) but it will not contain the updated rdbModel with connection retries since it has not yet been tested. It should take only another day for Warren to do necessary testing.

• Documentation: (Chuck) sends the following:

  Miscellaneous changes include adding a "Group Work Space" page and updating the "Individual User Work Space on SLAC Public" page, linking them in from the appropriate "splash" pages. I also removed all references to u31 and u33 and, thanks to Tom, added more detailed instructions on how to monitor resources to the "Using the SLAC Batch Farm" page.

  I've also posted a "blind" link to the new "gtgrb Tutorial" for Nicola and Richard to review and comment. I'm going to be adding additional explanatory notes to this page before making it public. In addition, I'm working on the companion "rmfit Tutorial", but have run into a couple of stumbling blocks with respect to creating the free IDL virtual machine upon which rmfit is built.

• Core talk: Heather is off to the wilds of Vermont, in search of H1N1 vaccine, so no Core Talk this week.

previous

minutes index

next

J. Bogart, Last Modified: 01-Jun-2010 15:46:51 -0700