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

    p_manage_02: Not enough space in TEMP

    • Article Type: General
    • Product: Aleph
    • Product Version: 18.01

    We're in the process of updating our authority records and recreating the bib headings on (our test server), and we've run into a TEMP tablespace problem with p_manage_02.

    A quick recap: All authority records loaded and indexed successfully in abc10 & abc12. I ran p_manage_102 in abc01 for umn12 successfully, then for abc10 also successfully.

    However, p_manage_02 is stuck in an uncertain state - and abc01 is still locked - with this error occurring at the point of creating abc01.z01_id:
    ERROR at line 1:
    ORA-01652: unable to extend temp segment by 128 in tablespace TEMP

    We had 8000MB allocated for TEMP when it happened, and have since increased that by another 4000MB. (Our DBAs don't set TEMP tablespaces to auto-extend.) However, we don't know if that's enough and would like some advice. Also, we've like to avoid running p_manage_102 (twice) again if possible.

    Is there some way we can recover from this problem gracefully and ensure that we have enough space allocated in TEMP?

    Looking at abc01_p_manage_02.01363 on /exlibris/aleph/u18_1/alephe/scratch, I see this:

    setenv p_active_library "ABC01"
    setenv p_delete_flag "0"
    setenv p_start_doc_number_x "000000000"
    setenv p_end_doc_number_x "999999999"
    setenv p_acc_group ""
    setenv p_no_process_x "16"
    setenv p_force_chk "Y"
    setenv p_duplicate_mode "Y" <-------------------

    As described in "How To Run Index Jobs", section 2.3 and KB 5709, the p_duplicate_mode should be "N".

    The problem which you are having is in the p_manage_02_x.cycles step. This is the p_duplicate_mode step.

    KB 5709 says: "Note: Running p_manage_02 with "p_duplicate_mode Y" ("Run in Duplicate mode") can result in such duplicates. This is similar to p_manage_17 in "A" mode. We recommend that you always specify p_duplicate_mode N."

    Since I think you will want to run it on Prod with "p_duplicate_mode N", I think it is better for you to run it on Test in this same fashion. That way a successful run on Test will give you the assurance that it will also work on Prod.

    Note: You will need to rerun p_manage_102 before the p_manage_02 rerun.

    • Article last edited: 10/8/2013