Related Action referring to Ghost Process Model version

Hello,

The only way that could happen is if that related action is triggering a process model that is running (not a new one).

When you publish a new process model, non-completed processes of old process models are still running. You should check

1) Create a new record and do the necessary steps to execute that related action

2) Check in the monitoring non-completed running instances of your related action. If that is the case, delete those instances.

Let me know if that hel

When you configure a process model for record related action, the process models latest published version will be picked to trigger an instance.

Is the process model published, without any errors? We often face issues, with deleting the process model versions. I would not recommend to delete the previous version, it will impact the parent processes that are using these.

Make sure that the latest version is in state "Published" not "Published disabled".

I assume that the PM configured as a related action is not a wrapper model which in turn calls a subprocess.

I understand the latest should be picked up. That's exactly my quesiton.

Now I did delete old older process models since I wanted to stop pulling my hair. I'm just learning right now, so there is no other Record referencing the process model. Status is indeed just published and not disabled. Also no subprocesses.

I've just been suffering from n00b syndrome where some basic things are working for me (like in the other thread I'm reading a CDT instance which will not get assigned to my pv! variable while process model throws no error). This is for the same process model.

Thanks for replying and improving my understanding. I'll keep at it and hope to reply with something different.

I have deleted all existing process instances. This happens when creating new instances. As unbelievable as it may sound...

I'm just going to start from scratch with new process model, record action, etc. for my sanity sake.

Thanks!

Hello,

I think conceptually there is something wrong of what you are doing.

You said that you introduced some debug variables to know which version of your process model you are running. For checking which version of the process model you are running, you should go to your application, select monitoring and check the version that the "Process Model" columns shows you.

That will help you to

1) check that the correct process model version is being picked up

2) See which values you variables are picking.

Additionally, if you have a process model with a start form, you can do "start process from debugging", that way you see better what is running.

I would not be removing and adding references and old versions of process models as it can leave you a mess behind. If you want to do further test, I would suggest to do a dummy interface with a constant that points to your process model and link to start the process model with a given parameters (for instance that way you can test with the values you should be giving from the related action)

Good luck!

There's a way you can try this without having to start from scratch with your process model.  From your current process model, select "File" --> "Save As", which will save a copy of the original as a separate / new PM.  Name it something noticeably different, then choose it as the related action PM instead of the original one, and see what happens when you try running it.  If it works well, you could just switch to using the new one as your main related action PM, and ditch the old one.

For sanity sake, how about creating a new PM that has nothing to do with any that have come before.  Let it be just start node and end node.  Configure the related action to run that PM and do nothing instead of whatever it was doing.  Confirm that you can at least do that.

Afterward, try reconfiguring your related action to use the new PM you've been trying to get it to run.

Because deleting in Appian is super weird, perhaps create a related action with a completely different name, then make sure the PM you're trying to run doesn't have the same name as anything you deleted.

It could be that Appian sees an object with the same name as a deleted previous object, and runs the previously deleted one because it goes by names for reasons.

It's stuff like this that gave colleagues of mine the best practice of never deleting anything in Appian ever.

Great advice. I thought the same thing. The behavior is consistent. I've done this 3 times since morning.

I created a dummy CDT, a dummy process with User Input Task, a dummy interface form. Saved. Published.

I debugged process model, went To Tempo Tasks, completed User Input Task. Looked in my database table for record. All good.

I went to process model. Pretended to change something. Literally, I checked the 

appianfreak said:

I'm seeing such inconsistencies with my Appian

By any chance are you running a local install that you installed yourself?  In the past I've seen cases where incomplete / incorrect configurations result in weird / inconsistent behavior on the front end.