Rebuild nozebra.pl
Назва
rebuild_nozebra.pl — сценарій переіндексування бібліографічних та віторитетних MARC-записів у БД Zebra.
Використовуйте це пакетне завдання, щоб переіндексувати усі бібліотечні або авторитетні записи у Вашій базі даних Koha.
Опис
Параметри:
-b
- — index bibliographic records
-a
- — index authority records
-daemon
- — Run in daemon mode. The program will loop checking
- for entries on the zebraqueue table, processing
- them incrementally if present, and then sleep
- for a few seconds before repeating the process
- Checking the zebraqueue table is done with a cheap
- SQL query. This allows for near realtime update of
- the zebra search index with low system overhead.
- Use -sleep to control the checking interval.
- Daemon mode implies -z, -a, -b. The program will
- refuse to start if options are present that do not
- make sense while running as an incremental update
- daemon (e.g. -r or -offset).
-sleep 10 — Seconds to sleep between checks of the zebraqueue
table in daemon mode. The default is 5 seconds.
-z — select only updated and deleted
records marked in the zebraqueue
table. Cannot be used with -r
or -s.
--skip-deletes — only select record updates, not record
deletions, to avoid potential excessive
I/O when zebraidx processes deletions.
If this option is used for normal indexing,
a cronjob should be set up to run
rebuild_zebra.pl -z without --skip-deletes
during off hours.
Only effective with -z.
-r — clear Zebra index before
adding records to index. Implies -w.
-d — Temporary directory for indexing.
If not specified, one is automatically
created. The export directory
is automatically deleted unless
you supply the -k switch.
-k — Do not delete export directory.
-s — Skip export. Used if you have
already exported the records
in a previous run.
-nosanitize — export biblio/authority records directly from DB marcxml
field without sanitizing records. It speed up
dump process but could fail if DB contains badly
encoded records. Works only with -x,
-w — skip shadow indexing for this batch
-y — do NOT clear zebraqueue after indexing; normally,
after doing batch indexing, zebraqueue should be
marked done for the affected record type(s) so that
a running zebraqueue_daemon doesn't try to reindex
the same records - specify -y to override this.
Cannot be used with -z.
-v — increase the amount of logging. Normally only
warnings and errors from the indexing are shown.
Use log level 2 (-v -v) to include all Zebra logs.
--length 1234 — how many biblio you want to export
--offset 1243 — offset you want to start to
example: --offset 500 --length=500 will result in a LIMIT 500,1000 (exporting 1000 records, starting by the 500th one
)
note that the numbers are NOT related to biblionumber, that's the intended behaviour.
--where — let you specify a WHERE query, like itemtype='BOOK'
or something like that
--run-as-root — explicitily allow script to run as 'root' user
--wait-for-lock — when not running in daemon mode, the default
behavior is to abort a rebuild if the rebuild
lock is busy. This option will cause the program
to wait for the lock to free and then continue
processing the rebuild request,
--table — specify a table (can be items, biblioitems, biblio, biblio_metadata) to retrieve biblionumber to index.
biblioitems is the default value.
--help or -h — show this message.