At 06:26 PM 1/5/2012 +1300, Mason James wrote:
On 2012-01-5, at 1:56 PM, Chris Cormack wrote: [snip] its skipping *all* records after. looks like you have a bad bib record
I'm guessing it aborted the process, as a quick look shows that none of the "missing" authorities have magically re-appeared.
How about trying it without the -b, just -a to reindex authorities. Id add -v for verbose too
yes indeed - there was an small bug with rebuild_zebra.pl script when indexing both bibs and auth records using the xml switch (not sure if its fixed yet?)
for best results, use the -x (for xml option) on authority records alone, like this...
$ ./bin/migration_tools/rebuild_zebra.pl -a -r -v -x
Thank you both very much. In fact the -a is incompatible with -x, but using -a -r -v, : 09:56:26-05/01 zebraidx(2947) [log] Iterations: isam/dict 885740/153588 09:56:26-05/01 zebraidx(2947) [log] Dict: inserts/updates/deletions: 153588/0/0 09:56:27-05/01 zebraidx(2947) [log] Records: 10398 i/u/d 10398/0/0 so the authorities are now (I *think* properly and completely) indexed. However, for the biblios, using only -b -r -v, the indexing failed after 12,835 records: 09:58:45-05/01 zebraidx(3048) [warn] MARC: Bad offsets in data. Skipping rest 09:58:45-05/01 zebraidx(3048) [warn] Record didn't contain match fields in (bib1,Local-number) 09:58:45-05/01 zebraidx(3048) [log] error grs.marcxml.record /tmp/5H7Hpo5SlB/biblio/exported_records 8741259 09:58:51-05/01 zebraidx(3048) [log] Iterations: isam/dict 12930648/373712 09:58:51-05/01 zebraidx(3048) [log] Dict: inserts/updates/deletions: 373712/0/0 09:58:57-05/01 zebraidx(3048) [log] Records: 12835 i/u/d 12835/0/0 yet using -x it re-indexed all 14,899 biblios 10:00:49-05/01 zebraidx(4857) [log] Iterations: isam/dict 16437721/481418 10:00:49-05/01 zebraidx(4857) [log] Dict: inserts/updates/deletions: 481418/0/0 10:00:55-05/01 zebraidx(4857) [log] Records: 14899 i/u/d 14899/0/0 QUESTION: Does this warning "Record didn't contain match fields in (bib1,Local-number)" have a specific cause that has been seen before? And is there a way in staff-client to find the record matching (presumably biblio #, but I can't find it even in "advanced search) 12,836? [Or is it easier to work directly from the mysql?] Again, many thanks - Paul