3

I have a question regarding the Workflow manager 1.0 for SharePoint 2013. Suddenly the CPU on the server is 100%, it seems that it is caused by corrupted workflow. There are two instances of this workflow constantly creating messages in event viewer. Even after deleting the items from sharepoint and even after unpublishing the workflow from list, these errors persists. See the log below. Is there a way how to get rid of this corrupted workflow instance?

ServiceBus threw MessageNotFound exception while trying to receive a deferred message. Deferred message context details: MessageId = 7448, ActionNamespace = Microsoft.Workflow.Dispatcher, ActionName = ReceiveNotification, InstanceId = {45fe9f19-4da0-4cc6-91fa-dcab6fa43549}, ResumptionKey = d94a8b19-31b7-4e08-9f12-51d2e8b3e977, MessageTraceId = {78160b7d-6c71-48dc-8276-e3d51bbf810c}

Improve this question
3
  • Do you installed WFM on the same SharePoint farm? what're the current specifications of your server?
    – Mohamed El-Qassas MVP ♦
    Commented Jun 26, 2017 at 15:30
  • I have the WFM server running for 3 years without any CPU utilization. It started after two WFs hang in "Started" status. On the SP server, the items are deleted and workflow is unpublished, but something for these two instances is still remaining on the WFM server. My setup is a SharePoint 2013SP1 5 nodes farm (2WFE, 2APP, 1BI) and a standalone WFM server.
    – Josef Svejda
    Commented Jun 27, 2017 at 16:09
  • I know we're not really supposed to ask for help in the answers section, but did you ever resolve this issue? I have a SharePoint 2013 farm experiencing the exact same symptoms and I cannot narrow down the issue. I can't find any workflows that are going off the rails.
    – Alex
    Commented Dec 29, 2020 at 21:39

1 Answer 1

Reset to default
0

Did you check which process is consuming more utilization? Before you run the workflow, enable Verboseex. After running the workflow when you observe the CPU utilization reaches to 95 - 100%, take a perfmon tool logs for that process. This should give a better picture of that consumption.

When you run that workflow on a particular list, check the list item count, item level permissions and see that they are within limitations.

Improve this answer
3
  • Thank you very much for the response. It is not caused by running any workflow. The workflow is already unpublished from the list. But it seems that two started instances of the workflow have not been completed and cause these deferred messages. And those messages are raising the CPU to 100%. I dont know how to stop them. Once the service bus is started on the WF server all 3 processes consumes the CPU (Service bus 40%, service host 20% and message broker 20%).
    – Josef Svejda
    Commented Jun 27, 2017 at 16:03
  • Take a Perfmon log, it will show you what operations those processes are performing.
    – Achiever SP
    Commented Jun 27, 2017 at 16:51
  • Hello, but I already know what is causing the problem, processes: Service bus, service host and message broker. I also know the source of problem, it reports me the event viewer, that there are deferred messages for 2 incomplete workflow instances (see the description in my post). What I don't know is that if it is possible to forcefully remove those instances from WFM or how to get rid of them. I have even tried to update WFM SQL DB WFInstanceManagement and set them as "Completed" but it had no effect.
    – Josef Svejda
    Commented Jun 29, 2017 at 8:43

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.