Zebra index problem after upgrade to 3.18
I upgraded a 3.10 package installation on Debian squeeze to 3.18.03 and now zebra will not reindex. Behaviour was exactly like Bug 13396/13404 I looked to change to DOM indexing but the wiki instructions were unclear. I added <zebra_bib_index_mode>grs1</zebra_bib_index_mode> to the koha.conf file but no progress. Could someone kindly suggest a next step please. Thank you, Clint. koha-rebuild-zebra -f -v gives: Zebra configuration information ================================ Zebra biblio directory = /var/lib/koha/koha/biblios Zebra authorities directory = /var/lib/koha/koha/authorities Koha directory = /usr/share/koha/intranet/cgi-bin Lockfile = /var/lock/zebra_koha_koha/rebuild/rebuild..LCK BIBLIONUMBER in : 999$c BIBLIOITEMNUMBER in : 999$d ================================ skipping authorities ==================== exporting biblio ==================== 1............................................................................... 101............................................................................. etc 5001............................................................................ 5101.............................................. Records exported: 5147 ==================== REINDEXING zebra ==================== 07:46:54-16/02 zebraidx(11623) [warn] 60:0:XML error: junk after document elemen t 07:46:54-16/02 zebraidx(11623) [warn] 1:0:XML error: not well-formed (invalid to ken) ==================== CLEANING ==================== Zebra configuration information ================================ Zebra biblio directory = /var/lib/koha/koha/biblios Zebra authorities directory = /var/lib/koha/koha/authorities Koha directory = /usr/share/koha/intranet/cgi-bin Lockfile = /var/lock/zebra_koha_koha/rebuild/rebuild..LCK BIBLIONUMBER in : 999$c BIBLIOITEMNUMBER in : 999$d ================================ ==================== exporting authority ==================== Records exported: 0 ==================== REINDEXING zebra ==================== skipping biblios ==================== CLEANING ====================
clint.deckard schreef op ma 16-02-2015 om 08:02 [+1300]:
I upgraded a 3.10 package installation on Debian squeeze to 3.18.03 and now zebra will not reindex.
Just base on what you're showing here, you might have some bad data:
07:46:54-16/02 zebraidx(11623) [warn] 60:0:XML error: junk after document elemen t 07:46:54-16/02 zebraidx(11623) [warn] 1:0:XML error: not well-formed (invalid to ken)
that's causing the process to abort. We've often found that as time went on, Koha got stricter about its data, so things that were fine in the past cause problems now (which is correct, but annoying.) I think it's a change since 3.10 that Koha now uses XML for indexing by default. Have a go running koha-rebuild-zebra with '-u', which will tell it to use USMARC. This might get things working now, and you can look into figuring out what record it is that is causing the problem. -- Robin Sheat Catalyst IT Ltd. ✆ +64 4 803 2204 GPG: 5FA7 4B49 1E4D CAA4 4C38 8505 77F5 B724 F871 3BDF
Hi! On 15 February 2015 at 20:02, clint.deckard <clint.deckard@frontiers.co.nz> wrote:
I upgraded a 3.10 package installation on Debian squeeze to 3.18.03 and now zebra will not reindex. Behaviour was exactly like Bug 13396/13404
I looked to change to DOM indexing but the wiki instructions were unclear. I added <zebra_bib_index_mode>grs1</zebra_bib_index_mode>
grs1 is the old mode. You want to replace that with: <zebra_bib_index_mode>dom</zebra_bib_index_mode> Best regards, Magnus Enger libriotech.no
Thank you for your reply Magnus. I found the immediate problem which was a line marcxml.recordType:grs.sgml in zebra.biblios.cfg was commented. Uncommenting it fixed the issue. I think that changing to DOM indexing is not as straightforward as changing <zebra_bib_index_mode>grs1</zebra_bib_index_mode> to <zebra_bib_index_mode>dom</zebra_bib_index_mode> and I was unable to follow the wiki instructions, so left it as is. Many thanks, Clint. Magnus Enger wrote:
Hi!
On 15 February 2015 at 20:02, clint.deckard <clint.deckard@frontiers.co.nz> wrote:
I upgraded a 3.10 package installation on Debian squeeze to 3.18.03 and now zebra will not reindex. Behaviour was exactly like Bug 13396/13404
I looked to change to DOM indexing but the wiki instructions were unclear. I added <zebra_bib_index_mode>grs1</zebra_bib_index_mode>
grs1 is the old mode. You want to replace that with:
<zebra_bib_index_mode>dom</zebra_bib_index_mode>
Best regards, Magnus Enger libriotech.no
participants (3)
-
clint.deckard -
Magnus Enger -
Robin Sheat