vRA request stuck in "In Progress" state but VM created and available in Managed Machines

[shauncw]

Hi Folks,

I'm hoping you can help me resolve a puzzle.

A user submitted a VM request which started a vRO workflow. Part of the workflow pauses release of the VM until it verifies that a service request is marked completed. While this was pending, the vRO and vRA servers were rebooted to resolve another issue. When the vRO interface came back up, there was a duplicate workflow for the same VM request. About 1 hour later, the service request was marked complete and original workflow completed. However the duplicate workflow did not complete and the vRA request still shows as "In Progress". The VM is available in the Managed Machines listing.

How can I either complete the duplicate workflow or remove it from the vRO workflow listing? And, why did vRO create a duplicate workflow after reboot?

Any information or suggestions would be welcomed.

[gopanaboena]

Hi

We also faced the same problem, at the time of training, might be it is client problem.

after some time automatically come to completed state

[shauncw]

One update to this question. BOTH vCO workflows did complete at about the same time. However the vRA request still shows as "In Progress".

Accordning to gopanaboena, the situation may correct itself over time. So, I'll update this issue at end of day today with any changes.

If anyone has any additional comments or suggestions please post here.

Thanks.

[SeanKohler]

Yeah... happens all the time.  We get requests that lose their backing and get stuck in "In Progress" on a daily basis.  We haven't been able to figure out why, but for us... it has been a regular occurrence, even without infrastructure reboots.

I have a support request in on it.  I am told that VMWare is aware this is happening.  (e.g. we are not the only ones)

There is a KB article that explains how to clean out the request from the DB.  The problem with this action for me is that the request was a VALID request, and a machine was built based off the request.  Deleting/Destroying the request isn't a great idea in my opinion.  I think it would be better if something came back through and at least set the request to anything other than "In Progress"  (for example: "Unknown" or "Lost Request Backing").  Then we have the request present in our history, but the state is at least changed out of a "In Progress" status which could be considered a *good* state for any condition other than it being stuck.

If you have support, I recommend you submit a ticket.  It adds firepower to getting faster resolution.