jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sébastien Launay <sebastien.lau...@anyware-tech.com>
Subject Re: clustering problems - Jackrabbit 1.4
Date Tue, 25 Nov 2008 08:50:49 GMT
Hi all,

Alexander Klimetschek a écrit :
> Hi,
>
> the problem is most likely the first error you get: URIIndex not
> valid? javax.jcr.NamespaceException: URI for index 11 not registered.
> The following errors seem to occur because after failing to read it,
> Jackrabbit tries to create that node
> (cafebabe-cafe-babe-cafe-babecafebabe is the fixed uuid of the root
> node), which gives the duplicate primary key error, because the node
> is actually present in the database, but could not be read due to the
> namespace problem.
>
> How did you create the second cluster? Did you copy the
> repository/namespaces and repository/nodetypes directories in the
> repository directory to a fresh new cluster node? The namespaces
> directory contains the namespace index, which must be the same on all
> nodes. Clustering synchronizes them, but if you start with an empty
> cluster node and your data contains custom namespaces (or nodetypes),
> it might be required to copy these directories.
>   
AFAIK, this is only required if you have a direct child or property with
a custom
namespace on the jcr:root node because it first load root node before
trying to
synchronize (which will declare this custom namespace).
This issue is already registered in JIRA:
https://issues.apache.org/jira/browse/JCR-1558
> Regards,
> Alex
--
Sébastien Launay

Mime
View raw message