You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@jackrabbit.apache.org by "Christopher M. Logan" <ch...@gmail.com> on 2010/01/14 23:38:37 UTC
Remote repository
In jackrabbit 1.6 and with moving to 2.0 in mind, what is the recommended approach to connecting to a remote repository that will preserve full functionality?
-Christopher
Sent via BlackBerry by AT&T
Re: where does jackrabbit store it's data?
Posted by Thomas Müller <th...@day.com>.
Hi,
> are there any recommendations on how to store the data?
> If yes, why? What are the advantages of having a almost 100% database storage?
> What are the advantages of storing anything into filesystem or why does it make sense to mix it?
See http://wiki.apache.org/jackrabbit/DataStore#File_Data_Store "It is
usually faster than the DbDataStore, and the preferred choice unless
you have strict operational reasons to put everything into a database.
"
Regards,
Thomas
AW: where does jackrabbit store it's data?
Posted by GUNACKER Simon <sg...@EUROFUNK.COM>.
Thank you Thomas,
are there any recommendations on how to store the data?
If yes, why? What are the advantages of having a almost 100% database storage?
What are the advantages of storing anything into filesystem or why does it make sense to mix it?
Regards, simon
> -----Ursprüngliche Nachricht-----
> Von: Thomas Müller [mailto:thomas.mueller@day.com]
> Gesendet: Dienstag, 19. Jänner 2010 17:16
> An: users@jackrabbit.apache.org
> Cc: bdelacretaz@apache.org
> Betreff: Re: where does jackrabbit store it's data?
>
> Hi,
>
> Jackrabbit can store (almost) everything in a database,
> everything on
> the file system, or a mix. It's can't store the Lucene
> index in a
> database. See also:
> http://wiki.apache.org/jackrabbit/ConfigurationOverview
>
> Regards,
> Thomas
>
>
> On Tue, Jan 19, 2010 at 5:09 PM, GUNACKER Simon
> <sg...@eurofunk.com> wrote:
> > hello Bertrand,
> >
> > thank you for your fast reply.
> >
> > The reason why I am asking is because we have been
> using oracle's content db till now.
> > It stores everything In database; binaries are stored
> as BLOB's.
> >
> > I recently found jackrabbit and I was quite amazed.
> > But my supervisor was asking me about how jackrabbit
> saves it's data.
> > I think he really likes everything being stored in
> database and he wanted to know
> > where jackrabbit stores its content.
> >
> > I said: I don't know - I think it's a mixture between
> database and filesystem and he asked
> > me about the advantages of this way.
> >
> > So, Bertrand what you are actually saying is that I
> could actually use any storage I like serving as
> backend?!
> > Am I understanding you correct that nodes and
> properties are stored in some database but file content
> is stored in filesystem?
> > Am I able to configure jackrabbit to store everything
> at one place?!
> >
> > Regards,
> > Simon
> >
> >> -----Ursprüngliche Nachricht-----
> >> Von: Bertrand Delacretaz
> >> [mailto:bdelacretaz@apache.org]
> >> Gesendet: Dienstag, 19. Jänner 2010 16:43
> >> An: users@jackrabbit.apache.org
> >> Betreff: Re: where does jackrabbit store it's data?
> >>
> >> Hi,
> >>
> >> On Tue, Jan 19, 2010 at 4:24 PM, GUNACKER Simon
> >> <sg...@eurofunk.com> wrote:
> >> > ...Where does jackrabbit actually store it's
> data?...
> >>
> >> Data storage is handled by persistence managers (for
> >> the nodes and
> >> properties) and if enabled by a file-based DataStore
> >> for large
> >> binaries. By default (IIRC) the standalone webapp's
> >> persistence
> >> manager is configured to use the Derby embedded
> >> database for storage.
> >> Other persistence managers are file-based, but
> anyway
> >> the storage
> >> mechanism isn't visible, nor relevant, from the JCR
> >> API's point of
> >> view.
> >>
> >> See "persistence managers" in
> >> http://jackrabbit.apache.org/frequently-asked-
> >> questions.html, and
> >> (optionally)
> >> http://wiki.apache.org/jackrabbit/DataStore
> >>
> >> > ...what's the difference to
> >> > using a 100% database solution?...
> >>
> >> In short, the JCR API provides high-level
> functionality
> >> for content
> >> management that relational databases don't. The
> >> http://wiki.apache.org/jackrabbit/JcrLinks page
> >> provides links to a
> >> series of articles that might be useful.
> >>
> >> -Bertrand
> >>
> >
> >
Re: where does jackrabbit store it's data?
Posted by Thomas Müller <th...@day.com>.
Hi,
Jackrabbit can store (almost) everything in a database, everything on
the file system, or a mix. It's can't store the Lucene index in a
database. See also:
http://wiki.apache.org/jackrabbit/ConfigurationOverview
Regards,
Thomas
On Tue, Jan 19, 2010 at 5:09 PM, GUNACKER Simon <sg...@eurofunk.com> wrote:
> hello Bertrand,
>
> thank you for your fast reply.
>
> The reason why I am asking is because we have been using oracle's content db till now.
> It stores everything In database; binaries are stored as BLOB's.
>
> I recently found jackrabbit and I was quite amazed.
> But my supervisor was asking me about how jackrabbit saves it's data.
> I think he really likes everything being stored in database and he wanted to know
> where jackrabbit stores its content.
>
> I said: I don't know - I think it’s a mixture between database and filesystem and he asked
> me about the advantages of this way.
>
> So, Bertrand what you are actually saying is that I could actually use any storage I like serving as backend?!
> Am I understanding you correct that nodes and properties are stored in some database but file content is stored in filesystem?
> Am I able to configure jackrabbit to store everything at one place?!
>
> Regards,
> Simon
>
>> -----Ursprüngliche Nachricht-----
>> Von: Bertrand Delacretaz
>> [mailto:bdelacretaz@apache.org]
>> Gesendet: Dienstag, 19. Jänner 2010 16:43
>> An: users@jackrabbit.apache.org
>> Betreff: Re: where does jackrabbit store it's data?
>>
>> Hi,
>>
>> On Tue, Jan 19, 2010 at 4:24 PM, GUNACKER Simon
>> <sg...@eurofunk.com> wrote:
>> > ...Where does jackrabbit actually store it's data?...
>>
>> Data storage is handled by persistence managers (for
>> the nodes and
>> properties) and if enabled by a file-based DataStore
>> for large
>> binaries. By default (IIRC) the standalone webapp's
>> persistence
>> manager is configured to use the Derby embedded
>> database for storage.
>> Other persistence managers are file-based, but anyway
>> the storage
>> mechanism isn't visible, nor relevant, from the JCR
>> API's point of
>> view.
>>
>> See "persistence managers" in
>> http://jackrabbit.apache.org/frequently-asked-
>> questions.html, and
>> (optionally)
>> http://wiki.apache.org/jackrabbit/DataStore
>>
>> > ...what's the difference to
>> > using a 100% database solution?...
>>
>> In short, the JCR API provides high-level functionality
>> for content
>> management that relational databases don't. The
>> http://wiki.apache.org/jackrabbit/JcrLinks page
>> provides links to a
>> series of articles that might be useful.
>>
>> -Bertrand
>>
>
>
AW: where does jackrabbit store it's data?
Posted by GUNACKER Simon <sg...@EUROFUNK.COM>.
hello Bertrand,
thank you for your fast reply.
The reason why I am asking is because we have been using oracle's content db till now.
It stores everything In database; binaries are stored as BLOB's.
I recently found jackrabbit and I was quite amazed.
But my supervisor was asking me about how jackrabbit saves it's data.
I think he really likes everything being stored in database and he wanted to know
where jackrabbit stores its content.
I said: I don't know - I think it’s a mixture between database and filesystem and he asked
me about the advantages of this way.
So, Bertrand what you are actually saying is that I could actually use any storage I like serving as backend?!
Am I understanding you correct that nodes and properties are stored in some database but file content is stored in filesystem?
Am I able to configure jackrabbit to store everything at one place?!
Regards,
Simon
> -----Ursprüngliche Nachricht-----
> Von: Bertrand Delacretaz
> [mailto:bdelacretaz@apache.org]
> Gesendet: Dienstag, 19. Jänner 2010 16:43
> An: users@jackrabbit.apache.org
> Betreff: Re: where does jackrabbit store it's data?
>
> Hi,
>
> On Tue, Jan 19, 2010 at 4:24 PM, GUNACKER Simon
> <sg...@eurofunk.com> wrote:
> > ...Where does jackrabbit actually store it's data?...
>
> Data storage is handled by persistence managers (for
> the nodes and
> properties) and if enabled by a file-based DataStore
> for large
> binaries. By default (IIRC) the standalone webapp's
> persistence
> manager is configured to use the Derby embedded
> database for storage.
> Other persistence managers are file-based, but anyway
> the storage
> mechanism isn't visible, nor relevant, from the JCR
> API's point of
> view.
>
> See "persistence managers" in
> http://jackrabbit.apache.org/frequently-asked-
> questions.html, and
> (optionally)
> http://wiki.apache.org/jackrabbit/DataStore
>
> > ...what's the difference to
> > using a 100% database solution?...
>
> In short, the JCR API provides high-level functionality
> for content
> management that relational databases don't. The
> http://wiki.apache.org/jackrabbit/JcrLinks page
> provides links to a
> series of articles that might be useful.
>
> -Bertrand
>
Re: where does jackrabbit store it's data?
Posted by Bertrand Delacretaz <bd...@apache.org>.
Hi,
On Tue, Jan 19, 2010 at 4:24 PM, GUNACKER Simon <sg...@eurofunk.com> wrote:
> ...Where does jackrabbit actually store it's data?...
Data storage is handled by persistence managers (for the nodes and
properties) and if enabled by a file-based DataStore for large
binaries. By default (IIRC) the standalone webapp's persistence
manager is configured to use the Derby embedded database for storage.
Other persistence managers are file-based, but anyway the storage
mechanism isn't visible, nor relevant, from the JCR API's point of
view.
See "persistence managers" in
http://jackrabbit.apache.org/frequently-asked-questions.html, and
(optionally) http://wiki.apache.org/jackrabbit/DataStore
> ...what's the difference to
> using a 100% database solution?...
In short, the JCR API provides high-level functionality for content
management that relational databases don't. The
http://wiki.apache.org/jackrabbit/JcrLinks page provides links to a
series of articles that might be useful.
-Bertrand
where does jackrabbit store it's data?
Posted by GUNACKER Simon <sg...@EUROFUNK.COM>.
hello there,
I'm new to jackrabbit and it's philosophy.
I have installed it, I populated it with some data but there is one
point I am missing:
Where does jackrabbit actually store it's data?
Is it 100% database or partly filesystem / what's the difference to
using a 100% database solution?
Please give me an idea on what this is all about.
thanks,
Simon
Re: Remote repository
Posted by ChadDavis <ch...@gmail.com>.
On Sun, Jan 17, 2010 at 11:49 AM, Christopher M. Logan
<ch...@gmail.com> wrote:
> I've been seeing other messages on the mailing list talking about the
> spi2jcr for remote access. Was the spi meant for this usage? Is there
> anything available that shows remote access using the spi?
> Should WebDAV be used over the spi?
>
SPI is a necessary part of the WebDAV remoting process, someone
correct me if I'm wrong. Here's the WebDAV JCR remoting architecture
as I understand it.
REPOSITORY SERVER:
The repository server side, as I call it, will use a servlet to expose
the repository for WebDAV remoting with a servlet,
org.apache.jackrabbit.j2ee.JcrRemotingServlet. This servlet can be
found in the jackrabbit-webapp maven module's WAR artifact. Run this
in a servlet container, etc.
CLIENT SIDE:
The client application connects with a stack of two components.
jcr2spi + spi2davex. Assembling this stack, in 2.0 at least, is
handled by using the appropriate factory to create your repository
object.
RE: Remote repository
Posted by "Christopher M. Logan" <ch...@gmail.com>.
I've been seeing other messages on the mailing list talking about the
spi2jcr for remote access. Was the spi meant for this usage? Is there
anything available that shows remote access using the spi?
Should WebDAV be used over the spi?
-Christopher
-----Original Message-----
From: Jukka Zitting [mailto:jukka.zitting@gmail.com]
Sent: Friday, January 15, 2010 7:58 AM
To: users@jackrabbit.apache.org; chrismikall@gmail.com
Subject: Re: Remote repository
Hi,
On Thu, Jan 14, 2010 at 11:38 PM, Christopher M. Logan
<ch...@gmail.com> wrote:
> In jackrabbit 1.6 and with moving to 2.0 in mind, what is the recommended
> approach to connecting to a remote repository that will preserve full
functionality?
Jackrabbit 2.0 will have two remoting mechanisms, RMI and WebDAV. From
a project roadmap perspective the WebDAV remoting is the preferred
alternative with the RMI layer kept mostly for backwards compatibility
and functional completeness.
The RMI layer was a bit troublesome to port to 2.0 but it now works.
The upgraded RMI layer is not yet available in the current 2.0 betas,
but will be included in the final Jackrabbit 2.0 release. The RMI
layer in Jackrabbit 2.0 will cover pretty much all of the JCR 1.0 API,
but not many of the JCR 2.0 extensions.
The WebDAV remoting layer has a better internal design than the RMI
layer and can thus support notably improved performance for many
operations. WebDAV remoting isn't as thoroughly tested as the RMI
layer and may lack some functionality, but I expect it to improve
quickly during the 2.x release cycle.
Both mechanisms implement the same JCR API, so it will not be too
troublesome to switch from implementation to another later on if
needed.
BR,
Jukka Zitting
Re: Remote repository
Posted by Jukka Zitting <ju...@gmail.com>.
Hi,
On Thu, Jan 14, 2010 at 11:38 PM, Christopher M. Logan
<ch...@gmail.com> wrote:
> In jackrabbit 1.6 and with moving to 2.0 in mind, what is the recommended
> approach to connecting to a remote repository that will preserve full functionality?
Jackrabbit 2.0 will have two remoting mechanisms, RMI and WebDAV. From
a project roadmap perspective the WebDAV remoting is the preferred
alternative with the RMI layer kept mostly for backwards compatibility
and functional completeness.
The RMI layer was a bit troublesome to port to 2.0 but it now works.
The upgraded RMI layer is not yet available in the current 2.0 betas,
but will be included in the final Jackrabbit 2.0 release. The RMI
layer in Jackrabbit 2.0 will cover pretty much all of the JCR 1.0 API,
but not many of the JCR 2.0 extensions.
The WebDAV remoting layer has a better internal design than the RMI
layer and can thus support notably improved performance for many
operations. WebDAV remoting isn't as thoroughly tested as the RMI
layer and may lack some functionality, but I expect it to improve
quickly during the 2.x release cycle.
Both mechanisms implement the same JCR API, so it will not be too
troublesome to switch from implementation to another later on if
needed.
BR,
Jukka Zitting