Process and human task variable constraints

[stmarci]

Greetings!

Our company is planning to utilise jBPM in one of out projects involving business processes. We decided to use jBPM 5.2, and our plan is to store large XML data in process variables (talking about a couple hundred megabytes) and mediate these from (human) task to (human) task and query these from a custom application through the Task Client API. These large files are supposted to be maps by the way and as such they could be even larger. Our 3rd party application would visualize these maps and present a view for clients based on process variables.

My question is, that is it possible for jBPM to store such large data in process variables, and whether are there any constraints regarding the storage limits for these. If you have any experience with such large process variables, I would like to hear your a priori experience regarding the performance impacts or other issues as well.

I appreciate your help and thank you for your time in advance,

Márton

[eaa]

Since variables are part of the process instance, they are persisted everytime the process instance is persisted.

So, if you are planning to use persistence, I wouldn't recommend you to store the whole xml inside the process. Is there any reason why you can't store just the id or name of the xml?

Best Regards,

[stmarci]

Thank you for your answer!

Storing the id is indeed a good idea! It also reminded me that runtime state persistence is not enabled by default in jBPM 5.

Best Regards,