the problem isn't that the redirection is not working but the cache generation is not proper terminated.
So in the cluster db (ezdfsfile), there is an entry of a cache file which doesn't exist. Apache is trying to get this file, but don't, sleeps one second and tries again. The apache process isn't terminated.
So, my question is: how to terminate ezpublish (after the redirection) in a proper way?
Georg, the problem happens because after sending the redirect header your http code exits, without giving eZP's cluster code a chance to finalize its cache file creation / cleanup.
If you do as Hamdi said, the node template will end correctly its execution, having a small file generated in the cache with only the definition of the persistent var in it. Then the pagelayout will do the header() call and terminate cleanly eZP execution. In short: it will work.
2nd thing: you're not running the latest version of eZ, are you? This misbehaviour of the caching layer has been fixed in enterprise patchsets and also in the latest community versions...
Principal Consultant International Business
Member of the Community Project Board
