What does it do to query performance ?

IIRC Compass has a JDBCDirectory that stores the indexes in DB blobs, but last 
time I looked (ok that was 3 years ago), the typical Lucene seek operation 
causes lots of network traffic especially where the DB didnt support seek into 
a blob. (at the time mysql didn't).

The other problem might be that Jackrabbit does some odd things with the index 
to update it without requiring readers to re-open, I cant remember what exactly.

Ian



On 16 Mar 2010, at 19:23, Joseph Ottinger wrote:

> Not that I know of. I'll see if I can hunt down my implementation and post
> it. It's really *lucene* and compass integration, where compass provides
> extra storage mechanisms for Lucene, but that should be available for
> Jackrabbit, too.
> 
> On Tue, Mar 16, 2010 at 3:14 PM, Justin Edelson 
> <[email protected]>wrote:
> 
>> Joseph-
>> Is Jackrabbit/Compass integration documented somewhere?
>> 
>> Thanks,
>> Justin
>> 
>> On 3/16/10 3:04 PM, Joseph Ottinger wrote:
>>> You can use a different Lucene storage mechanism, one that doesn't use a
>>> filesystem; I've used Compass' stuff with success.
>>> 
>>> On Tue, Mar 16, 2010 at 2:55 PM, Rakesh Vidyadharan <[email protected]
>>> wrote:
>>> 
>>>> 
>>>> On 16 Mar 2010, at 11:35, pradeepkudale wrote:
>>>> 
>>>>> 
>>>>> Hi all,
>>>>> 
>>>>> I was trying to configure Apache Jackrabbit 2.0.0 in order to get
>>>> everything
>>>>> stored in an mySql.
>>>>> It should not create any files on the LocalFileSystem but in my case it
>>>> is
>>>>> creating some files.
>>>> 
>>>> It is not possible to totally get rid of the local file system unless
>> you
>>>> totally disable search indexing.  I think even with that, JR will need
>> the
>>>> repository and workspaces directories on the local file system.
>>>> 
>>>> Rakesh
>>> 
>>> 
>>> 
>>> 
>> 
>> 
> 
> 
> -- 
> Joseph B. Ottinger
> http://enigmastation.com

Reply via email to