Can we have a 1TB jackrabbit repository?

[Stefan Kurla <st...@gmail.com>]

We plan on using jackrabbit for a repository that could easily go up
to 500GB and 1TB at the extreme. All the data will be in a single
workspace.

The workspace could be designed as a file system replacement with
workspace-related data stored in a database (DB2 or Oracle) and the
filesystem storing the indexes.

I have been trying to import the jackrabbit sources from svn (about
78MB) into the repository and I ran into problems which are explained
in a separate email. Hence I am concerned whether jackrabbit can
handle the system load that we are planning on.

Has anyone been able to scale to this kind of repository size? What
are the experiences?

S.

Re: Can we have a 1TB jackrabbit repository?

[David Nuescheler <da...@gmail.com>]

hi stefan,

we found in our tests that the overall size of the workspace does
not really have to much of an impact but rather the number of nodes
in the workspace.
we ran tests up to 120m nodes (60m files) without running into issues.
so i would argue that if your average node size is somewhere in 10k size
(or in our test: the average file size in the 20k) range you can easily
scale to 1tb per workspace.
i think it is important to choose the right persistence manager (a bundle
persistence manager, probably).
we ran our tests with a derby bundle persistsencemanager in its
default config with 128m heap size.

let me know if that answers your question.

regards,
david

On 4/25/07, Stefan Kurla <st...@gmail.com> wrote:
> We plan on using jackrabbit for a repository that could easily go up
> to 500GB and 1TB at the extreme. All the data will be in a single
> workspace.
>
> The workspace could be designed as a file system replacement with
> workspace-related data stored in a database (DB2 or Oracle) and the
> filesystem storing the indexes.
>
> I have been trying to import the jackrabbit sources from svn (about
> 78MB) into the repository and I ran into problems which are explained
> in a separate email. Hence I am concerned whether jackrabbit can
> handle the system load that we are planning on.
>
> Has anyone been able to scale to this kind of repository size? What
> are the experiences?
>
> S.
>