After task completion workflow still marked as running but doesn't continue to next activity

Question

We have a Visual Studio workflow for SharePoint 2013 running correctly several hundred times over the last few years. There are however a few workflow instances that do not continue when completing the related workflow task but are still marked as "running". We do not have access to the logs as the corresponding approval tasks were executed several months ago.

According to some blog entries this seems to happen from time to time (while most of the information is rather outdated), but until now we haven't found a convincing cause for it and the "solution" is usually to reinstall the workflow, the task list, etc.

In our case this is not an option as we're in a production environment and don't want to lose any data on document versions, approvers, etc. We would prefer to "resume" the flow from its current state if there is any way to do that.

When investigating this, we had a look at the involved database tables and stored procedures, especially the AllUserData table with the task item information as well as the Workflow table. There is an InstanceData field with a varbinary value (0x1234234234...) that might contain more information about the state of the workflow instance but we were unable to decrypt it for now.

Anybody knows what could be the cause for this behaviour and how we can resume the workflow? Is there a way to decrypt the InstanceData field in order to debug the workflow?

Answer 1

(sorry that this is marked as an answer but I miss 19 reputation to write a comment) Maybe you can describe your workflow in more detail to see if there's a common point in the cases the workflow's still in running state. My first guess would be, that one or more users/groups assigned to your workflow are deactivated or even deleted. In common the workflow should stop and fail if it's unable to select a pre-defined user/group. Therefore you have a VS Workflow I can't tell if there's any mistake by catching those errors.

Answer 2

One thing worth checking: I have seen some workflow instances get stuck in running mode because it was trying to send mails ( check that the SMTP does not have outages and if you can identify the actors of one instance that failed check that all the users involved have the mail property informed.

Other than that I am afraid that without being able to access the ULS log you are mostly shooting in the dark, if you can update the workflow you could try to add some debug messages to the history....

Answer 3

If only happened on some workflow instances, may be you should check your workflow design, it seems task completion failed to update the workflow status, maybe some condition changed cause the udpate workflow status failed or not happened, but no exception throwed.

You can check Workflow table, find the workflow id, and then check SchecduledWorkItems table with where id = [find workflow id], check whether there are some work item should be deliveried long time ago, but still in ScheduledWorkItems table. If the DeliveryDate field is before than current time, it means this work item wasn't be deliveried successfully.