Details
-
Bug
-
Resolution: Done
-
Critical
-
WCM 5.0 SP2
-
None
-
Sp2 17517
Description
On the 10'000 pages DB (on PG), I go into the su pages in my left menu (last item in my left menu) and then I simply launch the menu's engine on the left menu. Then I only click on next. This turns the workflow status of the page to red (why? nothing changed) and the save (or just the navigation in the engines' embedded navbar which automatically save the last content object) leads to really bad time (~15 sec to navigate from one object to the other...).
This looks like in jamon due to org.jahia.hibernate.manager.JahiaFieldsDataManager.preloadActiveOrStagedFieldsByMetadataOwner.
Moreover the indexing process, as the workflow state got changed, are also taking a lot of time. Each indexing order take nearly one minute to execute (and each time I navigate in my engine's navbar, this save the contaner and then launch a new indexing thread).
So:
- why are we saving and modifying the workflow state if the user did not change anything on the container (I even did not click on any other fields, the engine just opened on the title and I click on the next object arrow)? This will also generates ESI cache flushing... briefly speaking we should absolutely avoid saving containers which have not been changed. In such a case, this should really be only "read-only" access to such engines.
- Why does the indexing orders takes so much time to index... no change as nothing changed.
Why this org.jahia.hibernate.manager.JahiaFieldsDataManager.preloadActiveOrStagedFieldsByMetadataOwner takes so much time?
Extract of the console:
2007-06-11 15:21:20,234: INFO [org.jahia.services.search.lucene.AbstractLuceneSearchIndexer] - Finished processing 3 indexing orders in 71922ms.
2007-06-11 15:24:01,328: INFO [org.jahia.bin.Jahia] - Processed [/jahia/Jahia/engineName/updatecontainer/op/edit/pid/4179] user=[siteadmin] ip=[127.0.0.1] in [17375ms]