Last modified 13 years ago Last modified on 11/28/07 18:38:38

Delopers meeting Nov 28 2007

Participants : Timur, Tigran, Gerd, Dmitry, Patrick

  • gridKa lost > 13000 file in T0D1 after the files have been fetched the first time. After noticing that file are missing we immediately suggested to set all T0D1 pool 'permanent' which Doris did. The code fix is in 1.8.0-4 which is already installed at gridKa. SARA and NDGF have been informed immediately. No missing files have been reported anywhere else.
  • Space meeting at CERN. Only Altas has been present. Miguel B. provided a letter on how ATLAS intends to use Spaces.
  • Gerd's PinManager. Martin does a code review. How shell we proceed with the deployment.
    • Ask Lionel if PinManager still dies
    • Martin will become familiar with the PinManager and the SMC.
    • Timur will add more functionality to the new PinManager
  • Patrick makes sure, ATLAS understands the way dCache is currently handling spaces.
    • T1D0, and T1D1 restores go into non-managed space (not in link groups), which is difficult to understand for T1D1 in case a pool breaks.
  • Next dCache developer meeting in Iceland (3. week of Feb)
  • Patrick is on vacation from Nov 29 (Tomorrow) to Dec 20.
  • To serious problems detected at the CMS system.
    • Pools are selected for reading which are still in the process of restoring the file. This will be checked by Gerd. It is clear that we do this in 1.8 but not clear why it happens with 1.7.
    • Restore requests on the pools sometimes leave the internal file status in a bad state. (----S) This doesn't recover automatically even after the script dies.

Last modified by Patrick @ Wed Mar 3 07:18:39 2021