Erik Buene wrote:
to be the same on all cluster nodes. Only then the job mechanism is
able to detect that the same job has been started on different nodes.
Yes, this would be the next problem. For testing purposes I've been
making id's based on the node path, to ensure this would be the same on
all instances.
Ok, yes, creating this unique id can be a pain; with JCR 2.0 this will
be easier as there will be a time stamp of the change which should be
unique together with the path.
The custom handler on both server instances in the cluster (nodes in the
cluster) receives the event (the custom event, not the first
org/apache/sling/event/job), and on both instances there is no
application.id - which means both instances treats it as being a "local"
event, and starts to process it.
Argh, ok, then this is definitly a bug, could you try to debug the
problem? Or could you set the logger to level DEBUG and provide the logs?
Thanks
Carsten
--
Carsten Ziegeler
[EMAIL PROTECTED]