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

    ue_01 stops -- without any error message [z980, bad doc number problem]

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

    Description:
    Even though there are hundreds of abc01 z07 records waiting to be processed, ue_01 stops (that is, it stops writing to the ue_01 log and stops doing any processing), even though the three processes continue to show as running in util c/1. There is no error message.

    We have restarted it several times (-- needing to kill the ue_01_a process, since util e/2 doesn't stop it). Sometimes it will process two or three records but always quickly stops.

    The problem does not seem to be with a specific document: it will sometimes successfully process exactly the same doc# it stalled on before.

    Resolution:
    Like you, I don't find any consistency in the record being processed (the "HANDLING DOC NO"), but in each case the last line in the run_e_01 log is:

    Update z98 from z980 : 019000823129

    As seen in the abc01 tab00.eng, the "W-019" word index is "WCT" and util f/4 for the z97 shows that word# 000823129 is "abc01":

    enter start word position nnnnnnnnn
    000823129
    -abc01 -000823129-000823129

    So the problem is with the Word indexing of the word "abc01" from the CAT field.

    Since all of the records in the WCT index are for abc01 I don't think that it is useful to have "abc01" as a word in this index. I suggest *not* indexing the $$b (cataloger level) and $$l (library) subfields.

    To accomplish this, you would change the tab11_word from:

    ! 1 2 3 4 5 6 7 8 9 10 11 12
    !!!!!-!!!!!-!-!!!!!!!!!!-!!!!!!!!!!!!!!!!!!!!-!!-!-!-!!!!!!!!!!!!!!!!!!!!...

    CAT## 03 WCT

    to:

    ! 1 2 3 4 5 6 7 8 9 10 11 12
    !!!!!-!!!!!-!-!!!!!!!!!!-!!!!!!!!!!!!!!!!!!!!-!!-!-!-!!!!!!!!!!!!!!!!!!!!...

    CAT## -bl 03 WCT

    and then restart ue_01.

    It may be that ue_01 will just fail on some other tag, but this is the first case that I have seen or heard of where "Update z98 from z980" fails in this fashion, so I think that the change is definitely worth a try.

    After making this change, the site no longer consistently got the "Update z98 from z980 : 019000823129" message, but ue_01 still stalled.

    We found that there was a BIB doc with a wrong doc number: ‘?»¿000001’ . This apparently caused all the problems because of the z980 records (changes in Words) where it appeared. There were 5 of them, all of them for CAT index. We deleted them (as well as the doc itself) and now ue_01 is running smoothly.

    See also, KB 16384-24964.


    • Article last edited: 10/8/2013