Page 1 of 1

Search Module - crashes during reindexation ~43.413.600 words

Posted: Tue Dec 28, 2010 9:23 am
by waldemar_r
Hi there,

I've just created new module using CTLModuleMaker. I've imported more than 36000 records.

12 columns x 36000 rows = 432000 fields with TEXT data.
Each field consist of 5-100  ~ 43.413.600 words

After some time it crashes with 500 Server error. I have 46849 indexed words (each time at about the same amount).

Should I write some small search form with my own SQL Query? Or there is some solution to index this mount of text data?

Re: Search Module - crashes during reindexation ~43.413.600 words

Posted: Tue Dec 28, 2010 6:27 pm
by Dr.CSS
Please paste your server settings, you may be running out of memory...

Re: Search Module - crashes during reindexation ~43.413.600 words

Posted: Wed Dec 29, 2010 11:43 am
by waldemar_r
CMSMS  1.9.1

Configuration:

php_memory_limit:
255M

process_whole_template:
false

output_compression:
false

max_upload_size:
64000000

I've just noticed is that during indexation it went up to 49000 INDEXEs with 90000 Redords, after that it was "reset" to 0 and started to grow up to below values (stopped due to Internal Server Error)

cms_index_search_index INDEX 41808
Records 83 616

What for is the "cms_module_search_items" table? I can see that there are only 5100 records.
From here, the module takes content to index?

Re: Search Module - crashes during reindexation ~43.413.600 words

Posted: Wed Dec 29, 2010 4:33 pm
by Jeff
Pleas give us ALL the system settings. Including PHP timeout.

Re: Search Module - crashes during reindexation ~43.413.600 words

Posted: Wed Dec 29, 2010 4:58 pm
by waldemar_r
Cms Version: 1.9.1

Installed Modules:

   * CMSMailer: 2.0
   * FileManager: 1.0.3
   * MenuManager: 1.7.2
   * ModuleManager: 1.4
   * News: 2.11
   * nuSOAP: 1.0.2
   * Printing: 1.1.1
   * Search: 1.6.7
   * ThemeManager: 1.1.3
   * TinyMCE: 2.8.1
   * ModuleMaker: 0.3.2
   * CTLModuleMaker: 2.0.3
   * Katalog: 0.1


Config Information:

   * php_memory_limit: 255M
   * process_whole_template: false
   * output_compression: false
   * max_upload_size: 64000000
   * default_upload_permission: 664
   * url_rewriting: mod_rewrite
   * page_extension: .html
   * query_var: page
   * image_manipulation_prog: GD
   * auto_alias_content: true
   * locale:
   * default_encoding: utf-8
   * admin_encoding: utf-8
   * set_names: true


Php Information:

   * phpversion: 5.2.6-ogc1
   * md5_function: On (True)
   * gd_version: 2
   * tempnam_function: On (True)
   * magic_quotes_runtime: Off (False)
   * E_STRICT: 0
   * memory_limit: 255M
   * max_execution_time: 2000
   * output_buffering: 4096
   * safe_mode: Off (False)
   * file_uploads: On (True)
   * post_max_size: 20M
   * upload_max_filesize: 20M
   * session_save_path: /tmp (1777)
   * session_use_cookies: On (True)
   * xml_function: On (True)


Server Information:

   * Server Api: cgi-fcgi
   * Server Db Type: MySQL (mysql)
   * Server Db Version: 5.0.51a

-----

mysql.allow_persistent On On
mysql.connect_timeout 60 60

Re: Search Module - crashes during reindexation ~43.413.600

Posted: Wed Jan 05, 2011 10:07 pm
by waldemar_r
Some ideas?

Re: Search Module - crashes during reindexation ~43.413.600

Posted: Thu Jan 06, 2011 9:17 am
by fredp
waldemar_r wrote:Some ideas?
Hi,

I had similar Search module "reindex" crashes with a text-rich site, back on release 1.8.x. The site has lots of text: transcribed documents, census data in HTML tables, etc. Importing the pages worked fine, but reindex crashed. :( It appeared at the time that Search module reindexing didn't scale well for content such as this, but I didn't investigate the crashes much. As a short-term workaround, I marked the pages as "NotSearchable", in the page template:

Code: Select all

<!-- pageAttribute: NotSearchable --> 
That stopped the crashes, but the Search module became less effective. So, I temporarily switched to using an external search engine for searches. I'd like to revisit this sometime soon and get the Search module working for this site. If I find anything helpful, I'll post it here.

Hope this helped.

Re: Search Module - crashes during reindexation ~43.413.600

Posted: Thu Jan 06, 2011 4:28 pm
by waldemar_r
I don't understand why the reindexaion goes up to 89000 records (in cms_module_search_index) and is reset to 0. All records are deleted. Than the reindexation starts from where it finised (not from begining). What is wrong?


I extended the memory limit up to 2000MB so the script indexed more content (not 40000 but 90000 recs). Is there any solution to manage with this topic?