Skip to main content
  • Subscribe by RSS
  • Ex Libris Knowledge Center

    After running batch index jobs, ue_01 is reindexing the whole database

    • Article Type: General
    • Product: Aleph
    • Product Version: 16.02

    Description:
    I ran the re-indexing jobs on TEST to make sure I understood the process correctly. Apparently I don’t.

    The UE_01 process seems to be re-indexing the whole database and I don’t know what triggered that event. Based on the count in z07, it still has over 600,000 records to go so it is about half way through the process.

    Here is what I did:

    1. Ran p_manage_102.
    2. Ran p_manage_02 (However, I made a mistake here and selected rebuild instead of update so I think I wiped out the results of 102.)
    3. Started UE08 when 02 finished since I had wiped out the 102 results.
    4. Ran p_manage_105
    5. Ran p_manage 17
    6 Ran p_manage_35
    7. Ran P_manage_32
    8. Ran P_manage_01

    At this point, I noted that the system was very busy (uptime > 3.00) and started looking around and saw that the UE_01 log was growing. I checked the “count” in z07 and it was up over 1 million, about the size of the bib. file.

    What caused the system to think it had to reindex the whole database?

    Resolution:
    The z07 records are being generated by ue_08. When ue_08 is run with Create Z07 = Y, then a Z07 is generated for each bib record connected to the heading.

    The following grep in the vcc01 $data_scratch shows this:

    >grep -c 'Update doc -' *run_e_08.7*
    run_e_08.70402:3341030
    run_e_08.71874:0
    run_e_08.76754:79398
    run_e_08.76760:31335

    If you run p_manage_02 correctly (specifying Update, since it's been preceded by p_manage_102), you will not have this problem.

    I ran the following SQL which confirms that all of the Z07's have been generated by ue_08:

    SQL-ABC01> select * from z07 where z07_sequence like '199%';
    no rows selected

    (Since ue_01 processes the z07 records on a "first-in-first-out" basis, the online programs backdate the Z07's they create to 199n.)

    Therefore, you can do this on test to handle this:

    1. dlib abc01
    2. util e/10 to stop ue_08
    3. util a/17/1 for the z07 (to empty it out)
    4. util e/8 to start ue_08. In response to "Write Z07s for related bib records?" Type "N".

    See SKB 4093 for a general discussion of Z07 creation and processing.

    Additional Information

    faq, ue-01, ue01


    • Article last edited: 10/8/2013