On 06/07/2010, at 3:39 PM, gmoniey wrote:

> I've got the delayed deltas working on my machine (as well as upgraded
> sphinx/ts), but I had a couple of questions before I go to production.
> 
> 1. I've noticed that there is a ThinkingSphinx.remote_sphinx config
> field. I was wondering if this needs to be set to true for my setup (I
> have 2 site machines, one of which has searchd running on it, so only
> 1 machine is really remote). I thought that all that needed to be done
> was to address/port for the production section of sphinx.yml.

It's best to set remote_sphinx to true at the end of your environment.rb (with 
some logic for the environment) - the reason this is needed is that Thinking 
Sphinx looks for the pid file to confirm Sphinx is running - and when Sphinx is 
on a different machine, it can't do that. Setting this to true makes TS assume 
Sphinx is always running.

> 2. Are there any best practices for running the delayed_delta job?
> Cron, inittab, etc.

Maybe something like god or monit is a better fit? As it should be constantly 
running, instead of run at a regular interval (otherwise, cron would be a good 
fit).

> 3. Does the delayed delta job need to be restarted to retrieve updates
> (i.e. adding new searchable models).

Definitely a good idea to restart the queue after any structural changes.

Cheers

-- 
Pat


> 
> Thanks.
> 
> On Jun 29, 7:48 pm, Pat Allan <[email protected]> wrote:
>> In Delayed Job 1.8.5, it looks it checks for jobs every 5 seconds, but you 
>> can change that:
>> 
>>   Delayed::Worker.sleep_delay = 2
>> 
>> I assume it's the same in the 2.x releases (when I last checked, I don't 
>> think the code had changed that dramatically).
>> 
>> --
>> Pat
>> 
>> On 30/06/2010, at 12:41 PM, gmoniey wrote:
>> 
>> 
>> 
>>> I see...I guess my only option is to go with delayed jobs. From what I
>>> have read in the docs, the delayed_delta job is a long living process
>>> which periodically checks for delta objects. Do you have any details
>>> on how often it wakes up and processes deltas? Is possible to set this
>>> field?
>> 
>>> Thanks.
>> 
>>> arash
>> 
>>> On Jun 29, 6:35 pm, Pat Allan <[email protected]> wrote:
>>>> Yup, that's exactly the problem you'd be facing.
>> 
>>>> --
>>>> Pat
>> 
>>>> On 30/06/2010, at 11:30 AM, gmoniey wrote:
>> 
>>>>> Hi Pat,
>> 
>>>>> Thanks for responding. I would like to stay away from delayed deltas
>>>>> as we need items to be added to the indexes in realtime. If the index
>>>>> files are accessible by all servers, is the issue then contention over
>>>>> the same files when attempting to update the deltas?
>> 
>>>>> On Jun 29, 5:55 pm, Pat Allan <[email protected]> wrote:
>>>>>> When you're dealing with multiple servers, the best delta approach is 
>>>>>> delayed deltas - and so you have the delayed job queue running on the 
>>>>>> same machine that your index files (and probably searchd) are located on.
>> 
>>>>>> That way, all the Sphinx files and processes are in one place, and delta 
>>>>>> requests get put into the database, and then are handled by the queue.
>> 
>>>>>> --
>>>>>> Pat
>> 
>>>>>> On 30/06/2010, at 10:45 AM, gmoniey wrote:
>> 
>>>>>>> I'm transitioning my app to a multiple server setup, and I am trying
>>>>>>> to figure out what the best approach is. From what I gather, the
>>>>>>> following should be fine:
>> 
>>>>>>> 1. Run searchd on 1 machine, and have it point to indexes on a nas
>>>>>>> (which is accesible by all machines)
>>>>>>> 2. Point other machines to the 1 searchd instance.
>> 
>>>>>>> My concern/confusion is with deltas; from what I have read on the
>>>>>>> other posts, the delta indexing may fail, but I'm not quite sure why.
>>>>>>> Is it that multiple apps may be trying to write to the same files on
>>>>>>> the nas?
>> 
>>>>>>> --
>>>>>>> You received this message because you are subscribed to the Google 
>>>>>>> Groups "Thinking Sphinx" group.
>>>>>>> To post to this group, send email to [email protected].
>>>>>>> To unsubscribe from this group, send email to 
>>>>>>> [email protected].
>>>>>>> For more options, visit this group 
>>>>>>> athttp://groups.google.com/group/thinking-sphinx?hl=en.
>> 
>>>>> --
>>>>> You received this message because you are subscribed to the Google Groups 
>>>>> "Thinking Sphinx" group.
>>>>> To post to this group, send email to [email protected].
>>>>> To unsubscribe from this group, send email to 
>>>>> [email protected].
>>>>> For more options, visit this group 
>>>>> athttp://groups.google.com/group/thinking-sphinx?hl=en.
>> 
>>> --
>>> You received this message because you are subscribed to the Google Groups 
>>> "Thinking Sphinx" group.
>>> To post to this group, send email to [email protected].
>>> To unsubscribe from this group, send email to 
>>> [email protected].
>>> For more options, visit this group 
>>> athttp://groups.google.com/group/thinking-sphinx?hl=en.
> 
> -- 
> You received this message because you are subscribed to the Google Groups 
> "Thinking Sphinx" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to 
> [email protected].
> For more options, visit this group at 
> http://groups.google.com/group/thinking-sphinx?hl=en.
> 

-- 
You received this message because you are subscribed to the Google Groups 
"Thinking Sphinx" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/thinking-sphinx?hl=en.

Reply via email to