I have a customer in financial activity and they are migrating their processes from Adobe Central Server to LiveCycle Output ES4. For mass document generation, at the moment they have designed two processes which take an xml from the watched folder, then a pdf is generated as result of merging the xml with an XDP template (the pdf document has only one page), and finally the pdfs are sent via email and are stored in system directory.
In a test with the processes designed in LiveCycle Workbench setting up a watched folder (the above mentioned), they sent 2000 xml for one process and they got an outcome of 2000 pdfs in 7 minutes and if the second process is test in a simultaneous manner with the first process, they get a time of 13 minutes to complete 4000 xmls and generate 4000 pdfs ( 2000 xmls for each LC process ).
My costumer is complaining about this ratio. They have the same processes in Adobe Central Server and they can get 2000 pdfs in 3 minutes. Also they have some processes that attend requests on demand, and due to performance issue, the request is stacked so they get a delay in their web portal (the LiveCycle process is invoked from a web portal and a pdf document is shown).
The LiveCycle Server environment is following:
Red Hat Enterprise Linux Server 6.0
WebLogic 10.3.6 (WLS Start Mode=Production)
Oracle 11g (LiveCycle database is hosted in a remote server)
RAM: 8 Gb
Running in a virtual machine (2 core processors (AMD 64)).
JVM HotSpot 64-bits Server
oMinimum Heap: 1085Mb
oMinimum Heap: 3640Mb
Default POOL_MAX (=4) for XMLForm.exe
My customer is planning to migrate an estimate of 400 jobs to LiveCycle Output ES4 so they will get a workload of 150,000 documents per day. How can I achieve best performance in LiveCycle Output?