[Koha-devel] [Koha] Zebra Indexing Problems

[Chris Mueller]

Environment: Koha 3 Alpha / Ubuntu 7.1.0 / Zebra 2.0.24 ** was Zebra 
2.0.26 ** / Mac-Tel Hardware / MARC21 

Some of our problems decribed in original message below have been 
resolved by replacing Zebra 2.0.26 with Zebra 2.0.24 -- OPAC searching 
works again. John Chadwick, our Information Technology Manager, has 
reported the Zebra 2.0.26 problem as Bugzilla Bug #1811. 

Answers etc. to Joshua's questions in reply to my original message: 

-- Zebraqueue daemon running 
-- Make target used = standard 

-- Biblio that was still indexed after deletion from database had been 
deleted using the "staff client" API [biblioitem deleted first, then 
biblio]; this problem cleared when we reverted to Zebra 2.0.24 and 
rebuilt the zebra indexes -- but, now when we attempt to delete 
different biblioitem and biblio from database (using API) the "delete" 
options are grayed-out. 

-- After reverting to Zebra 2.0.24 we still have a delay in indexing. 
What is norm time between adding a new biblio and it being indexed, 
without our having to rebuild the index? 

-- We also had a problem with deleting a patron record using the "staff 
client" API. We could not delete record using MS Internet Explorer, but 
were able to delete it using Firefox. 

-- More details on loss of directories when server is rebooted. This is 
method used to recreate directories which are wiped out: 

In the directory /var/run there needs to exist the following directories 
(owner and group KOHA). So: 

sudo mkdir koha 
sudo chown koha koha 
sudo chgrp koha koha 
cd koha 

Repeat the above process only substitute zebradb 

sudo chmod 777 * 

Now, cd /var/lock 

sudo mkdir koha 
sudo chown koha koha 
sudo chgrp koha koha 
cd koha 

Repeat the above process only substitute zebradb 

Now, make two directories that will be empty. One is biblios and one is 
authorities. 

sudo chmod 777 * 
........................ 
Any suggestions on how to improve speed of indexing and how to reboot 
server without loosing directories greatly appreciated! 

Christine Mueller 
Database Administrator 
Office of Information Technology 
New Mexico State Library 
1209 Camino Carlos Rey 
Santa Fe, NM 87507 
Phone: 505-476-9744 
christine.mueller at ... 
http://www.stlib.state.nm.us/

====================================================== 
Re: [Koha] Zebra Indexing Problems 

by Joshua Ferraro-3 Jan 31, 2008; 08:02am 

Hi Christine, 

On 1/30/08, Mueller, Christine, DCA <Christine.Mueller at ...> wrote: 
> 
> 
> Environment: Koha 3 Alpha / Ubuntu 7.1.0 / Zebra 2.0.26 / Mac-Tel 
Hardware / 
> MARC21 
> 
> No experience with Koha 2.x; our goal is to become familiar with Koha 
while 
> waiting for Koha 3 Beta 
Good plan :-) 

> Problems with Zebra Indexing: 
> 
> 
> 
> 1)       When biblio and biblioitem deleted from database data 
remained in 
> Zebra indexes and search results displayed link to biblio which was no 
> longer in database, even after indexes rebuilt. - How to prevent? 
Could you clarify how you deleted data from the database? Did you use 
the API, 
or directly query the RDBMS? 

Do you have the zebraqueue daemon running? 

> 
> 
> 2)       Sometimes new biblios would not be indexed for 20-30 minutes 
and 
> other times Zebra indexes would need to be rebuilt to force indexing 
of new 
> biblios. - Any suggestions for solving delayed indexing problem? 
Do you have the zebraqueue daemon running? 

> 3)       When Zebra server is restarted Zebra directories are missing 
and 
> must be restored. - How can server be restarted without deleting 
> directories? 
Hmm ... that's very strange ... which make target did you use? standard, 
dev? 

> 4)       Today searching no longer works at all. Zebra server was 
restarted, 
> missing directories were restored and indexes were rebuilt. Biblios 
and 
> biblioitems are still in the database. - Any suggestions for restoring 
> search function *without* reinstalling Zebra? 
i think we need to figure out why your zebra directories are 
disappearing ... 

> 5)       If we want to export the biblios and biblioitems (less than 
25 at 
> this time) in the database to create a backup file in case we need to 
do a 
> total reinstall of Koha and Zebra are there any "gotchas" or best 
practices? 
The MARC data is stored as a blob in biblioitems, so you could just 
mysqldump 
that table ... there are also some utilities in the interface in Tools 
-> Export Bibliographic and Holdings Data; as well as some command 
line tools that do the same thing in misc/ 
directory. 

So ... we need to figure out why your zebra configuration isn't working 
well, 
could you tell us a bit more about your environment? 

Also, I'd suggest we take this discussion to the koha-devel list where 
more of the 
technical questions tend to get hashed out (I've CCed your question to 
that 
list). Here's the list info: 

http://lists.nongnu.org/mailman/listinfo/koha-devel

Cheers, 

-- 
Joshua Ferraro                       SUPPORT FOR OPEN-SOURCE SOFTWARE 
President, Technology       migration, training, maintenance, support 
LibLime                                Featuring Koha Open-Source ILS 
jmf at ... |Full Demos at http://liblime.com/koha |1(888)KohaILS


-- 
View this message in context: http://www.nabble.com/Re%3A--Koha--Zebra-Indexing-Problems-tp15206767p15278739.html
Sent from the Koha - Dev mailing list archive at Nabble.com.