ue_01 Word indexing extremely slow
- Article Type: General
- Product: Aleph
- Product Version: 20
Problem Symptoms:
The xxx01 run_e_01_word.24779 log shows the following:
HANDLING DOC NO. - XXX01.007286081 2013-02-06 14:55:11
Update z98 from z980 : 028000002686
Update z98 from z980 : 016000011324
Update z98 from z980 : 026000015273
Update z98 from z980 : 027000529291
Update z98 from z980 : 016000636269
Update z98 from z980 : 037000673531
Update z98 from z980 : 026000712437
Update z98 from z980 : 026000771905
Update z98 from z980 : 001000803572
Update z98 from z980 : 001001985422
Update z98 from z980 : 001011771447
SQL shows that there are 5,400 z07a records waiting to be processed:
xxx01@ALEPH20> select count(*) from z07a;
**** Hit return to continue ****
COUNT(*)
----------
5436
At the rate of 3 records/minute it will take 30 hours to process these 5,400 z07a's.
Cause:
The slowness has to do with the bib records which ue_01 is processing. Looking at record 7286081, for instance, we see that it is an online resource, with 5xx fields and with an 856 field -- whose subfields u and y are being sent to the WRD and WUR word indexes.
The most common cause of slowness is large records with large 5xx note fields containing many words.
Resolution:
Such loads of online-resource bib records -- or bib records with large 5xx fields -- are probably best done over a weekend.
Additional Information
The following Articles are at least somewhat relevant:
000014414 (ue_01 slow indexing Early English Books Online)
000010926 (Preventing/reducing "Update z98 from z980" ue_01 messages)
000016804 (ue_01 hanging processing EEBO records)
000016316 (Word indexing of 856 subfield u (URL) doesn't work)
Category: Background processing (500)
Subject: Ue (500)
- Article last edited: 2/12/2015