I am having some problems installing version 3 of Koha, specifically, make test fails with a billion errors. I have downloaded the distribution. I created a database called koha (overwriting my version 2.x koha database). I made sure the kohaadmin user had access to the database. I ran perl Makefile.PL and kept all the defaults. When I run make test I get bunches o' errors mostly because the database has no tables. Was the make process expected to create the database structure? Did I miss something? -- Eric Lease Morgan
Hi, On 2/24/08, Eric Lease Morgan <eric_morgan@infomotions.com> wrote:
Was the make process expected to create the database structure? Did I miss something?
The database structure is created by the web installer, which is invoked the first time that you access the staff web interface (use the mysql database username and password to log in). The test cases have been revised to not attempt the DB-dependent tests during installation -- this change is available via a git pull of the current development version of Koha 3 and will be available in the beta distribution as soon as it is released. Regards, Galen -- Galen Charlton Koha Application Developer LibLime galen.charlton@liblime.com p: 1-888-564-2457 x709
On Feb 24, 2008, at 4:24 PM, Galen Charlton wrote:
Was the make process expected to create the database structure? Did I miss something?
The database structure is created by the web installer, which is invoked the first time that you access the staff web interface (use the mysql database username and password to log in).
Thank you for the prompt rely, but now I need some hints batch loading into version 3.0 I have (more or less) deleted my version 2.x of Koha and re-started with version 3.0. I need to support millions of records, and everybody desires faceted browse. I have installed version 3.0 using the single user model. I have edited the koha-httpd.conf file so Apache can find my installation. I can connect to both the administrative and users interfaces. I tried using the Web interface to stage and add 100 MARC records. Staging them seems to go just fine but completing the staging seems to fall into an infinite loop. I quit the process after about twenty minutes while my browser's status bar tried to load hundreds of items. The % value never went above 0. I then got Zebra up an running. I tried to connect to it using yaz- client, but I was never successful. I tried some queries from the admin interface but nothing was ever returned, yet the zebra server logged some interactions. I tried using the commit_biblios_file.pl script. It found my staged data, but again nothing seemed to get imported. I then tried to use the bulkmarcimport.pl script. It counted my 100 test records but they never seemed to get added to the system. Again, what am I doing wrong? Have I not configured something? How can I ultimately load my millions of records into version 3.x? Finally, I think the version 3.x demo is really impressive! I like the path this is taking. Faceted browse. Cover art. Web Services interfaces. Very nice. The only thing I didn't see was spell check/ Did You Mean? -- Eric Lease Morgan
Hi, On 2/24/08, Eric Lease Morgan <eric_morgan@infomotions.com> wrot
Again, what am I doing wrong? Have I not configured something? How can I ultimately load my millions of records into version 3.x?
For handling millions of records, bulkmarcimport.pl is the better bet. The staging MARC import (for which I've fixed a few bugs recently -- are you using the alpha distribution or a recent git pull?) is handy if you want to try different record matching profiles or revert batch loads, but consumes extra space and time that's not needed for a load of millions of records. Once you do a bulkmarcimport.pl run, the zebraqueue daemon process will index the records one at a time. But if you've just done a large bib load, it's much faster to (re)build the indexes in bulk by running rebuild_zebra.pl. If you haven't started zebraqueue or run rebuild_zebra.pl yet, then Zebra hasn't indexed anything yet, and that might account for the lack of search results. If running rebuild_zebra.pl doesn't take care of it, please post some more details from the Zebra log.
Finally, I think the version 3.x demo is really impressive! I like the path this is taking. Faceted browse. Cover art. Web Services interfaces. Very nice. The only thing I didn't see was spell check/ Did You Mean?
The feature is there, but for minor typos, the search engine automatically includes the correct terms. For example, try searching on 'Stephen Wing'. It is also possible to configure Koha to explicitly give spelling suggestions to the user, but that doesn't appear to have been activated for opacdemo.koha.org. Regards, Galen -- Galen Charlton Koha Application Developer LibLime galen.charlton@liblime.com p: 1-888-564-2457 x709
participants (2)
-
Eric Lease Morgan -
Galen Charlton