Switching 2.1.0 RAR to PostgreSQLPersistenceManager

[Robert Campbell <rr...@gmail.com>]

Right now we're using a default RAR distribution in JBoss which our
webapp connects to over JNDI. When the repository is booted for the
first time, it creates a default repository.xml, which contains:

    <FileSystem class="org.apache.jackrabbit.core.fs.local.LocalFileSystem">
        <param name="path" value="${rep.home}/repository"/>
    </FileSystem>

in the Repository, Versioning, and Workspace tags. There is also:

        <PersistenceManager
class="org.apache.jackrabbit.core.persistence.pool.DerbyPersistenceManager">
          <param name="url" value="jdbc:derby:${wsp.home}/db;create=true"/>
          <param name="schemaObjectPrefix" value="${wsp.name}_"/>
        </PersistenceManager>

in the Versioning and Workspace tags, and:

   <DataStore class="org.apache.jackrabbit.core.data.FileDataStore"/>

in the Repository tag.

It's not clear to me how these entries interact, even after reading
the Wiki article on persistence. Do they mean that Jackrabbit persists
_everything_ to Derby, which then uses the local file system? Or - as
I suspect - does it mean that actual binary files/blobs are stored as
files (FileDataStore) on the local file system, but that version and
workspace data is stored in Derby?

I'm trying to understand step by step how we can move _everything_
over into Postgres. We're doing this because our operations folks have
heavily invested in PG knowledge/infrastructure/backup systems/etc.
We'd like the ability to flip the switch on our Postgres backup
solutions and have our entire repo backed up live (no downtime to the
app).

Thank you,

Rob Campbell

Re: Switching 2.1.0 RAR to PostgreSQLPersistenceManager

[Robert Campbell <rr...@gmail.com>]

Sorry, I forgot to include my plan (for easier feedback). From the
wiki documentation, it sounds like I want to:

1) Replace all instances of DerbyPersistenceManager with
PostgreSQLPersistenceManager.
2) Leave LocalFileSystem alone; we're running nix in Production, so it
should be fast enough. We don't need any special fs/disks for search
indexes
3) Delete the "<DataStore
class="org.apache.jackrabbit.core.data.FileDataStore"/>" config, so
that Jackrabbit defaults large BLOBs to the PersistenceManager (PG) so
we have one point of backup.



On Mon, Aug 9, 2010 at 4:31 PM, Robert Campbell <rr...@gmail.com> wrote:
> Right now we're using a default RAR distribution in JBoss which our
> webapp connects to over JNDI. When the repository is booted for the
> first time, it creates a default repository.xml, which contains:
>
>    <FileSystem class="org.apache.jackrabbit.core.fs.local.LocalFileSystem">
>        <param name="path" value="${rep.home}/repository"/>
>    </FileSystem>
>
> in the Repository, Versioning, and Workspace tags. There is also:
>
>        <PersistenceManager
> class="org.apache.jackrabbit.core.persistence.pool.DerbyPersistenceManager">
>          <param name="url" value="jdbc:derby:${wsp.home}/db;create=true"/>
>          <param name="schemaObjectPrefix" value="${wsp.name}_"/>
>        </PersistenceManager>
>
> in the Versioning and Workspace tags, and:
>
>   <DataStore class="org.apache.jackrabbit.core.data.FileDataStore"/>
>
> in the Repository tag.
>
> It's not clear to me how these entries interact, even after reading
> the Wiki article on persistence. Do they mean that Jackrabbit persists
> _everything_ to Derby, which then uses the local file system? Or - as
> I suspect - does it mean that actual binary files/blobs are stored as
> files (FileDataStore) on the local file system, but that version and
> workspace data is stored in Derby?
>
> I'm trying to understand step by step how we can move _everything_
> over into Postgres. We're doing this because our operations folks have
> heavily invested in PG knowledge/infrastructure/backup systems/etc.
> We'd like the ability to flip the switch on our Postgres backup
> solutions and have our entire repo backed up live (no downtime to the
> app).
>
> Thank you,
>
> Rob Campbell
>