[Mondrian] Catalog clustering

[Hilario Fernandes]

Thank you for all the replies!

So, either way ( saving on a web server or implementing a DSP to read from
database ) the whole catalog would be massive if i had hundreds of cubes.
That would generate a huge overhead when adding a cube to a catalog,
considering the whole catalog would have to be generated and the checksum
matched, when compared to the current RolapSchema.addCube method.

Am I wrong with this theory?

On Thu, Dec 4, 2014 at 8:22 PM, Julian Hyde <julianhyde at gmail.com> wrote:

> The simplest thing is to put the schemas in a web server. (Apache VFS
> generalizes this notion to other protocols… but HTTP is best IMHO.)
>
> On Dec 4, 2014, at 11:32 AM, Luc Boudreau <lucboudreau at gmail.com> wrote:
>
> To refresh the schema cache, you would need to ping each node.
>
> Roland has a good alternative, but the DSP is mondrian specific.
>
> My personal favorite is to use a custom Apache VFS provider. You tell
> mondrian to fetch the schema from an arbitrary URL, something like "
> myschemas://schema_v1.xml". On the flip side, you create a VFS handler
> which serves the "myschemas://" protocol. The only change required for
> Mondrian is to enable UseContentChecksum=true in your connection URL. This
> way the contents of the XML virtual file is the cache key, not it's URL.
>
> The downside is that you start leaving behind a lot of stale schemas. They
> eventually will go away when the GC frees up that space, but it's important
> not to create millions and millions of different schemas with tiny
> differences. As much as possible, try to use the same schema structure and
> leverage role based security whenever possible.
>
> On Thu, Dec 4, 2014 at 2:17 PM, Roland Bouman <roland.bouman at gmail.com>
> wrote:
>
>> This may be way off.
>>
>> But since your database is shared already, could you make an
>> implementation of the dynamic schema processor that reads the info for the
>> schema from the database?
>>
>> On Thu, Dec 4, 2014 at 8:11 PM, Hilario Fernandes <
>> hilario.fernandes at cortex-intelligence.com> wrote:
>>
>>>
>>> Hi,
>>>
>>> I'm working on an application that uses mondrian as it's Olap Server.
>>> The application can run several instances in a cluster environment sharing
>>> the same database.
>>>
>>> The problem is that the cube structure might change trough one of the
>>> instances, leaving the others with an outdated schema.
>>>
>>> Is there any kind o mechanism for storing the catalog that can be used
>>> to centralize it? Something similar to the cache, that can be implemented
>>> to our needs?
>>>
>>> Any pointers here from the community?
>>>
>>> Thanks
>>>
>>> --
>>> Hilario Fernandes
>>>
>>> _______________________________________________
>>> Mondrian mailing list
>>> Mondrian at pentaho.org
>>> http://lists.pentaho.org/mailman/listinfo/mondrian
>>>
>>>
>>
>>
>> --
>> Roland Bouman
>> blog: http://rpbouman.blogspot.com/
>> twitter: @rolandbouman
>> linkedin: http://www.linkedin.com/profile/view?id=5142800&trk=tab_pro
>>
>> Author of "Pentaho Solutions" (Wiley, ISBN: 978-0-470-48432-6
>> http://tinyurl.com/lvxa88) and "Pentaho Kettle Solutions" (Wiley, ISBN:
>> 978-0-470-63517-9 http://tinyurl.com/33r7a8m)
>>
>> _______________________________________________
>> Mondrian mailing list
>> Mondrian at pentaho.org
>> http://lists.pentaho.org/mailman/listinfo/mondrian
>>
>>
> _______________________________________________
> Mondrian mailing list
> Mondrian at pentaho.org
> http://lists.pentaho.org/mailman/listinfo/mondrian
>
>
>
> _______________________________________________
> Mondrian mailing list
> Mondrian at pentaho.org
> http://lists.pentaho.org/mailman/listinfo/mondrian
>
>


-- 
Hilario Fernandes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.pentaho.org/pipermail/mondrian/attachments/20141209/1e559265/attachment-0001.html