jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Colin Anderson" <colin.w.ander...@gmail.com>
Subject Re: Jackrabbit Clustering
Date Fri, 30 May 2008 10:37:15 GMT
On Fri, May 30, 2008 at 11:27 AM, Alexander Klimetschek
<aklimets@day.com> wrote:
> Sorry, did not see that you already switched to a shared lucene index config.

No worries :-)

>
> On Fri, May 30, 2008 at 12:14 PM, Colin Anderson
> <colin.w.anderson@gmail.com> wrote:
>> Thanks for replying. Can you give me some more detail as I'm not sure
>> what you mean? We are already using the BundleFSPersistenceManager for
>> the workspaces.
>
> You need a persistence manager that supports clustering. This is
> currently the case when you use a bundle db persistence manager
> configured to run on a clustered database. The actual data-clustering
> thus happens in the database. Jackrabbit's journal is used for
> synchronizing the order of events and Jackrabbit's high-level caches.
>
> The BundleFSPersistenceManager does not know about the clustering and
> thus both nodes write to the same files at the same time, which does
> not work.

Okay, that goes well with our next step.

Incidentally, could you use the BundleFSPersistenceManager if you used
an atomic filesystem?

Cheers,
Colin.

Mime
View raw message