Hi Thibault

I'm guessing you're using Capistrano? Best to make sure you've got 
configuration_file set in your thinking_sphinx.yml file to ensure that 
generated conf file lives in the shared directory. See the first example here:
http://pat.github.io/thinking-sphinx/deployment.html

(That setting was only added to the docs a few weeks ago, hence it may have 
been missed).

Once it's there, best to SSH in, kill any searchd processes, then run 
ts:rebuild (via cap or directly) and see how that goes.

-- 
Pat

On 6 Feb 2014, at 10:16 pm, Thibault Clavel <[email protected]> wrote:

> Oh it seems that there is no production.sphinx.conf in my config folder, 
> maybe that's why it does not seem to start at all. How can I generate this 
> one ?
> 
> 
> 
> Le jeudi 6 février 2014 11:52:18 UTC+1, Thibault Clavel a écrit :
> Thank you Pat, I found that solution in an old message of this board, that's 
> why I had deleted my last post.
> 
> The :source => :query method works well on localhost but I tested it 
> yesterday evening on my small dedicated server and it took the whole night 
> without finishing, not even telling me "indexing card_core".
> 
> Is there a way to launch rake ts:rebuild RAILS_ENV=production in verbose mode 
> to see what is happening ?
> 
> Thanks a lot.
> 
> Le mardi 4 février 2014 23:57:57 UTC+1, Pat Allan a écrit :
> There's no direct option for > or < - the easiest way to take that approach 
> is to use ranges, with the non-specific values really big or really small. 
> So, time greater than now could be :time => Time.zone.now..1.year.from_now 
> (or whatever's appropriate for your dataset).
> 
> On 5 Feb 2014, at 2:52 am, Thibault Clavel <[email protected]> wrote:
> 
>> Well, indexing time went from 20+ hours to less than a second, that's pretty 
>> amazing !
>> 
>> While I am at it, is there a way to use the with: parameters with < and > 
>> operators ? I haven't seen this in the documentation...
>> 
>> 
>> 
>> Le mardi 4 février 2014 11:56:14 UTC+1, Pat Allan a écrit :
>> :source => :query shifts the attribute into its own SQL statement, which is 
>> simpler and thus faster (I'm always surprised at how much faster it can be), 
>> rather than adding extra joins and complexity to the main query covers the 
>> rest of the index data.
>> 
>> Great to know it's helped!
>> 
>> -- 
>> Pat
>> 
>> On 4 Feb 2014, at 9:22 pm, Thibault Clavel <[email protected]> wrote:
>> 
>>> Hi Pat,
>>> 
>>>  Adding :source => :query worked like a charm, I tested with other 
>>> attributes thaht made indexing last forever and indexing is now lightning 
>>> fast !
>>> 
>>>  What did that parameter change ?
>>> 
>>>  Thanks a lot anyways for your quick and effective answer, I was stuck on 
>>> this since last week. :)
>>> 
>>> 
>>> Le mardi 4 février 2014 11:01:01 UTC+1, Pat Allan a écrit :
>>> Hi Thibault
>>> 
>>> Can you add the following option to each of the attributes you've defined: 
>>> :source => :query - and see if that helps?
>>> 
>>> Also: in TS v3, delta options are specified alongside :with => 
>>> :active_record:
>>> 
>>>   ThinkingSphinx::Index.define :card, :with => :active_record, :delta => 
>>> true do
>>>     # ...
>>>   end
>>> 
>>> -- 
>>> Pat
>>> 
>>> On 4 Feb 2014, at 8:11 pm, Thibault Clavel <[email protected]> wrote:
>>> 
>>>> Hi everyone,
>>>> 
>>>> I've set up TS 3.1.0 on a rails 4 app, which is what I would call a 
>>>> "playing card manager".
>>>> 
>>>> Here is my Card index :
>>>> 
>>>> ThinkingSphinx::Index.define :card, :with => :active_record do
>>>>   # fields
>>>>   indexes name, :sortable => true
>>>>   indexes name_en, :sortable => true
>>>> 
>>>>   has colors(:id), :as => :color_ids
>>>>   has card_formats(:id), :as => :card_format_ids
>>>>   has rarities(:id), :as => :rarity_ids
>>>> 
>>>> 
>>>>   set_property :delta => true
>>>> end
>>>> 
>>>> 
>>>> Colors, Card Formats and Raritiers are all has_many through relationships 
>>>> with the Card model. When I run rake ts:rebuild or ts;index, it takes less 
>>>> than a minute to proceed.
>>>> 
>>>> If I add this line :
>>>> 
>>>>   has card_types(:id), :as => :card_type_ids
>>>> 
>>>> ... which is another has_many :through, indexing starts ("indexing 
>>>> card_core") but seems to get stuck at this point. Last stime I let it run 
>>>> for almost 20h with no progress. :(
>>>> 
>>>> If I run SHOW INDEX on the join table card_card_types, I get this :
>>>> 
>>>> mysql> SHOW INDEX FROM card_card_types;
>>>> +-----------------+------------+---------------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
>>>> | Table           | Non_unique | Key_name                              | 
>>>> Seq_in_index | Column_name  | Collation | Cardinality | Sub_part | Packed 
>>>> | Null | Index_type | Comment | Index_comment |
>>>> +-----------------+------------+---------------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
>>>> | card_card_types |          0 | PRIMARY                               |   
>>>>          1 | id           | A         |       25550 |     NULL | NULL   |  
>>>>     | BTREE      |         |               |
>>>> | card_card_types |          1 | index_card_card_types_on_card_id      |   
>>>>          1 | card_id      | A         |       25550 |     NULL | NULL   | 
>>>> YES  | BTREE      |         |               |
>>>> | card_card_types |          1 | index_card_card_types_on_card_type_id |   
>>>>          1 | card_type_id | A         |          34 |     NULL | NULL   | 
>>>> YES  | BTREE      |         |               |
>>>> +-----------------+------------+---------------------------------------+--------------+--------------+-----------+-------------+----------+--------+------+------------+---------+---------------+
>>>> 
>>>> which seems OK to me, since card_id and card_type_id are indexes.
>>>> 
>>>> Any thoughts about this ?
>>>> 
>>>> Thanks a lot !
>>>> 
>>>> 
>>>> 
>>>> -- 
>>>> You received this message because you are subscribed to the Google Groups 
>>>> "Thinking Sphinx" group.
>>>> To unsubscribe from this group and stop receiving emails from it, send an 
>>>> email to [email protected].
>>>> To post to this group, send email to [email protected].
>>>> Visit this group at http://groups.google.com/group/thinking-sphinx.
>>>> For more options, visit https://groups.google.com/groups/opt_out.
>>> 
>>> 
>>> -- 
>>> You received this message because you are subscribed to the Google Groups 
>>> "Thinking Sphinx" group.
>>> To unsubscribe from this group and stop receiving emails from it, send an 
>>> email to [email protected].
>>> To post to this group, send email to [email protected].
>>> Visit this group at http://groups.google.com/group/thinking-sphinx.
>>> For more options, visit https://groups.google.com/groups/opt_out.
>> 
>> 
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "Thinking Sphinx" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to [email protected].
>> To post to this group, send email to [email protected].
>> Visit this group at http://groups.google.com/group/thinking-sphinx.
>> For more options, visit https://groups.google.com/groups/opt_out.
> 
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Thinking Sphinx" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to [email protected].
> To post to this group, send email to [email protected].
> Visit this group at http://groups.google.com/group/thinking-sphinx.
> For more options, visit https://groups.google.com/groups/opt_out.

-- 
You received this message because you are subscribed to the Google Groups 
"Thinking Sphinx" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/thinking-sphinx.
For more options, visit https://groups.google.com/groups/opt_out.

Reply via email to