Comments (2)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 85BE_Yogesh_Aggarwal commented Permalink

Hi Sergii

This is indeed a good article on points to be considered before migrating in-flight process instances to the new snapshot. I checked the product documentation and tried my luck on google but could not find an explicit statement on migration of failed process instance. As it is not documented anywhere, I believe that it is not supported by the product but there seem to be few real-time scenarios where it could fine use. Consider the following scenario
1) A human task is submitted successfully by the User 1.
2) A System Task follows the Human Task. This fails due to a code related issue (I know it should not happen in production env but the possibility can not be completely ruled out)
3) A human activity follows the System Task to be completed by User 2.
Now, User 1 has completed his task and User 2 will never get the task in his bucket. In addition, since the task is failed, code fixes done for to resolve the problem will also not be able to revive the failed process instance (since failed tasks do not get migrated).

2 SergeiM commented Permalink

Hi Yogesh,

Thank you for your comment.
It's absolutely possible to migrated Failed instances and this is one of the good examples where migrate option might be a good choice.
Example -
You have snapshot1 in your runtime environment and it has an error in the code causing instances to Fail.
Let's say you have 10 Failed instances in Snapshot1 in Runtime env.
You have created a fix in Process Center and made a new named snapshot - snapshot2.
Now if you try to install snapshot2 to your Runtime server and choose migrate option it would migrate all the Failed instances from previous snapshot1 to new snapshot2 code and you will be able to resume failed instances and they would resume from latest save execution point. (you can resume them using the API or Process Inspector).
Let me know if you have any questions.