jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sergiy Shyrkov <my.public.mail...@gmail.com>
Subject Re: Problem with Clustering Setup (PostgreSQL)
Date Thu, 28 Oct 2010 14:11:19 GMT
Hello Thomas,

thank you for the clarifications!

In that sense it means that using DatabaseFileSystem implementations in 
cluster produces a lot of configuration overhead as for each cluster node
(and each workspace + versioning) we need to manually create a separate 
DB schema or at least separate DB tables.
"Manually" in our case is unfortunately a consequence of the technical 
(security) limitations of our customers.
They run an application with a DB user, who has no permissions to create 
tables on-the-fly (during runtime).
That is why the Jackrabbit DB schema is created once by a DB administrator.

Could you, please, give me some hints if there are any advantages in 
using DatabaseFileSystem compared to LocalFileSystem?
As in our case it seems the LocalFileSystem is more appropriate solution 
at least from cluster configuration point of view.

I assume, the information about clustering configuration 
(http://wiki.apache.org/jackrabbit/Clustering ) is also valid for 
Jackarabbit 2.1.x/2.2.x. Is it correct?

Thank you in advance!

Kind Regards
Sergiy

On 28.10.2010 15:04, Thomas Müller wrote:
> Hi,
>
>>> Each cluster node needs its own (private) workspace level and version
>>> FileSystem (only those within the workspace and versioning configuration;
>>> the ones in the repository.xml and workspace.xml file).
>> Does it mean that the stuff under<Versioning>  must not reside in the DB?
> No, it means each cluster node needs its own FileSystem. That can be a
> DbFileSystem, but it needs to be its own (not a shared one).
>
>>         <FileSystem class="org.apache.jackrabbit.core.fs.db.DbFileSystem">
>>             <param name="driver" value="org.postgresql.Driver" />
>>             <param name="url" value="${jackrabbit.repository.url}" />
>>             <param name="user" value="${jackrabbit.repository.user}" />
>>             <param name="password" value="${jackrabbit.repository.password}"
> This is a shared one. Its own (private) means use a different database
> (or schema) for each cluster node, for example:
>
>          <FileSystem class="org.apache.jackrabbit.core.fs.db.DbFileSystem">
>              <param name="driver" value="org.postgresql.Driver" />
>              <param name="url"
> value="${jackrabbit.repository.url}${clusterNodeId}" />
>              <param name="user" value="${jackrabbit.repository.user}" />
>              <param name="password" value="${jackrabbit.repository.password}"
>
> Node the added "${clusterNodeId}".
>
> Regards,
> Thomas


Mime
View raw message