Skip to main content
ExLibris
  • Subscribe by RSS
  • ExLibris Dev

    Diagnosing ue_21 performance problems

    • Article Type: General
    • Product: Aleph
    • Product Version: 20, 21, 22, 23

    Description:
    Certain records are being processed by the abc01 ue_01 but no z00p is created: there are 290,000+ records in the abc01 z07p.

    ./abc01/scratch/run_e_21 is working, but extremely slowly (1 record every 20 seconds or so). The lmn01 ue_21 is taking 1/10th of a second per record. (The lmn01 ue_21 is running 200 times as fast.)

    Stopping/restarting the abc01 ue_21 does not help.

    Resolution:
    The 1/10th of a second per record is similar to what I see for the usm01 ue_21 on our server.

    1. Are there multiple publishing sets (multiple lines) specified in the abc01 tab_publish? If so, temporarily change it so just one publishing set (one line) is specified, then stop/start ue_21, and check the ue_21 performance in the run_e_21 log.

    If this improves performance, then try it with just the other dataset. And compare the abc01 tab_publish to the lmn01.

     

    2. Do some of the publishing sets have a non-standard base specified in column 2 of their line in tab_publish (-- a base other than the regular xxx01pub OPAC  base)?  If so, do some of those bases have truncated search strings (using the ? or * wildcard characters).  If so, test these searches in the OPAC to see if they are slow.  Truncated search values can slow down ue_21 processing.  This is discussed in the article " Logical base response very slow ".
     


    • Article last edited: 25-May-2016
    //doorbell.io feedback widged