• Main INDEX
  • Monthly INDEX
  • PREV
  • NEXT
    Make New Entry, Make Followup Entry

    User name R. Michaels

    Log entry time 17:00:08 on March21,2012

    Entry number 367238

    keyword=re: duplicate CODA files

    The version of CODA we're using for the 2 HRS has a problem that
    with ~2% probability the CODA file will have the run number of
    the previous file. First, I'll write some updated info on this
    problem, and then I'll write a "HOWTO" for shift workers to react.

    Since the 2nd file gets written to a next disk, we don't over-write
    the file, but the 2nd file has an identical name, so it won't
    go into MSS. As things are set up now, I must recover this by
    hand, which is risky and inconvenient. There are files in the
    MSS with the string "2nd" in them, e.g. /mss/halla/g2p/raw/g2p_3249_2nd.dat.0

    Whether this is run 3249 or 3250 depends on which file got
    written to MSS first. One may discover the run number from
    the CODA data structure.

    The DAQ group says that by looking at the printout from the
    Event Recorder (ER), i.e. coda_er, we may be able to understand
    the problem. We may also avoid the problem, probably, by not
    using the script that picks the next disk, instead writing to
    a link and changing this link by hand or by a cron script.
    We can make that change if the expt decides too, but it would
    help the long-term health of CODA to diagnose and fix this bug.

    --------------- HOWTO for Shift Workers -----------------

    Shift workers should do the following. The Run Coordinator may
    modify this procedure.

    In the paper run sheets, note the run number and that it has
    incremented by 1. Note the file name and that it's run number
    is the same.

    If "yes", proceed with the run.

    If "no", do the following:

    1. Make a screen capture in halog of the ER component window.

    2. Do a "startcoda" to start CODA over again.

    3. Do NOT delete any CODA files, even if you don't understand why.