harvester2 installtion issues

[shaoyusu]

Hi,

I just installed harvester2 on Ubuntu Dapper Drake (6.06 LTS) with Apache2 (2.0.55), MySQL 5.0.22, PHP 5.1.2. The installation finished without error message. But I meet the following issues:

1. If search has no result, will get error message
Fatal error: Call to a member function displaySummary() on a non-object in /var/www/harvester-2.0.0/cache/t_compile/%%92^924^92435639%%results.tpl.php on line 101
2. If search returns more than one record, only the first record is displayed; and the total record number shows up correctly. For example, if seach result has 40 records, the first page only shows the first record. Going to the next page (itermsPerPage is 25), also only record 26 is displayed. I can dump the $results array and it looks fine.
3. Adding more than one archive, only the latest one shows up.
4. After adding one archive, do “update metadata index” for it, only the first set of ListRecords is retrieved. It seems it doesn’t continue when a resumptionToken is met.
5. Trying to run command
php tools/upgrade.php upgrade
The command returns right away without doing anything, and no output, even no error message.

I guess issues 1, 2 and 3 have same problem since they all use Smarty’s iterator tag. But I check the Smarty installation, everything looks fine.

Any help will be appreciated!

Su

[asmecher]

Hi Su,

Hopefully this will correct a few items on your list:

The Smarty iterator problems can be corrected with a Smarty upgrade -- apply the patch at http://pkp.sfu.ca/bugzilla/show_bug.cgi?id=2393 and see if that corrects the problem.

For the command-line tools, double-check to see that your command-line php.ini is configured like your web-server's php.ini to include things like the MySQL driver. Many distributions use different configuration files (e.g. /etc/php5/cli/php.ini vs. /etc/php5/apache/php.ini).

Let me know how these go and I'll move on to the rest.

Regards,
Alec Smecher
Public Knowledge Project Team

[shaoyusu]

Alec,

Thanks for your reply. After applying the patch for smarty, record listing is fixed. Also, as you pointed out, I need to use
php -c /etc/php5/apache2/php.ini tools/harvest.php 1 flush

to force php cli to use ini file from apache2 setup. So issues 1,2,3,5 are all fixed! -:)

Issue 4 (only fecthing the first set of records) is still there. Using the command line harvesting has the same result:
=========================
dpubs@su-ubuntu:/var/www/harvester-2.0.0$ php -c /etc/php5/apache2/php.ini tools/harvest.php 1 flush
PHP Warning: Module 'mysql' already loaded in Unknown on line 0
Selected archive: Marriott Library Special Collection
Flushing metadata index for archive... 50 records deleted.
Fetching records...
Finished; 50 records indexed in 9 seconds (5.56 records per second).
=========================

It seems that although handleResumptionToken() in OAIHarvester.inc.php is called, it never gets into call back block since 'call back' is not set.

Thanks a lot!

Su

[asmecher]

Hi Su,

I'm unable to reproduce this last situation with PHP 4.4 or PHP 5.1.6, and I'm assuming you have more than 50 records in your repository. If you've reached a dead end on this issue, would you be willing to give me temporary access to your server to investigate? If so, please contact me directly. It might also help to send me the OAI URL you're using so I can see if it's a protocol issue.

Regards,
Alec Smecher
Public Knowledge Project Team

[shaoyusu]

Alec,

Please check your inbox. I sent you a private message.

Thanks,

Su

[shaoyusu]

Sorry, I forgot to include the OAI server: http://db2-sql.staff.library.utah.edu:8 ... ne/oai.jsp

[shaoyusu]

Alec pointed out that my OAI provider encodes the resumptionToken string. That conflicts OAI harvester soec and breaks the harvesting flow.

After fixing that, everything looks good. The harvesting is kind of slow though. My box is a P4 1.03G, only for testing.

Thanks a million!

Su

[asmecher]

Hi Su,

We haven't focused on improving the performance of harvesting, but I hope to spend time on this in the future. In the meantime, if you come across any ways of optimizing harvesting, please post them here for other users.

Regards,
Alec Smecher
Public Knowledge Project Team