I 've just found a workaround. I have mounted the four used tmp folders to /tmpfs (virtual memory). This means on author instance: tomcat/temp and webapp/author/tmp. Now the big activations finish successfully (also when started in parallel). I'm not sure what the real problem is. We uses GlusterFS as file storage system. But all other services runs well on that cluster file system. It looks more like a thread / sort order problem which is avoided by the very fast /tmpfs now.
About your questions: - yes, it was more-less random - it happens more on whole subtrees, but on small pages as well - yes, we uses workflow Regards, Michael -- Context is everything: http://forum.magnolia-cms.com/forum/thread.html?threadId=af4976dd-9af9-4b15-b0ec-562b2e034a15 ---------------------------------------------------------------- For list details, see http://www.magnolia-cms.com/community/mailing-lists.html Alternatively, use our forums: http://forum.magnolia-cms.com/ To unsubscribe, E-mail to: <[email protected]> ----------------------------------------------------------------
