This is indeed a known bug but not sure if someone has created an SR for this.
If you want to have it fixed I would recommend creating a service request for this.
Are you aware that there is a workaround for this where you bind the data source to a method in the managed bean and you build the query in java. This of course removes the runtime feature of changing the data source but at least it will work.
Thanks for your attention. In fact, I can use other way to get my result with CP by using RIDC service in java bean with complex queries. But if I do that, page load will longer (the node be got twice). If on a page we have >10 Content presenter. You can not imagine the page load time.
With Oracle AS Portal page load time so good(average 1,5 s). But with WebCenter Portal page load time so bad(average 5s)
In your case you can check/set following configuration for increase your performance:
- Configure JoC.
- Configure Coherence for CP caching in case of use ADF CP and not SS Region Templates. In case of Region Template, look at Oracle Support where you can find some optimizations in case of use Region Template.
- Configure WebCenter Content to prevent zip/unzip resources in the requests.
- Add lazy loading trick of A-TEAM blog post if you don't want to wait for the last TF load.
I've pages with 8 Content Presenter and they load ~1,5 - 2 secs.
Anyway is not recommedable have more than 5 TF in a page.
Thanks for your info. I will check and try to do with your suggestion
Configure WebCenter Content to prevent zip/unzip resources in the requests". I can't find out that.
Thanks a lot!
Is just an UCM configuration to prevent images to be zip/unzip with the request.
Add it to your UCM configuration:
<Sep 17, 2013 4:59:17 PM ICT> <Error> <oracle.as.cache.groupv2.HttpHandler> <BEA-000000> <CO-0 Invalid HTTP request information: GET /ubck/faces/vi/vimenu/vipages_trangchu HTTP/1.1>
I always generate the files using WLST but modifying it manually depending of the topology of the infrastructure.
Check if the listener and all addresses are OK. Also check during the restart of the manage server (in -diagnistic.log) if the JoC is initiated correctly or if it's having problems to initiate it.