Generically Handling Stuck Folders

Occasionally folders can become "stuck" in a stage if a flag isn't fired, conditionals or timers don't execute properly, etc.  It is easy enough on a one off basis to put in actions to allow an administrative user to move the folder appropriately, but when you are in a shared services environment, this becomes uneccessarily duplicative, so I'm thinking of creating my own generic process for handling these cases.  Has anyone else already done this?  Jerome - I'm talking to you.  :-)

 

I don't feel like wasting the effort if someone has already created something I could use. 

Tagged:

Comments

  • No, sorry Rob, we have not done anything like that.

     

    I have added the 'add flags' form to the admin tools to enable the adding of eWait entries for existing folders at a stage. That is generally how I go about it, although it is rather a 'manual' process. At least you do not have to invoke an action for each folder.

  • We have the same issue as well.  In most of our processes we use a common stage called refresh and a loop back action called refresh.  This is a one off fix as you described - but it seems (and we have not done this) we can use this pattern (refresh action) with the ECL.  Basically, write an app (web service or whatever) that queries to see which folders are stuck (amount of days sitting) then fire off the refresh action per folder using the ECL.  I would schedule this as a windows scheduler job or sql server job (maybe using ssis).

     

    This functionality of having a fail safe mechanism for stuck folders should be part of the enviornment and not our job to work around - imo.

     

  • Mike - You are on a similar line of thought to what I was thinking.  However, I was thinking of using an internal MBPM "Service" process for handling the checking. 

     

    I was actually thinking that it would use a "Refresh" loopback along with a "MoveTo" auto-forward action.  Solutions would register to the handler to specify a stage that they want to monitor and what mechanism a folder would be considered "stuck" and how it would be handled.  I think doing this would capture a few more use cases.

     

    I definitely agree that in an ideal world it would be OOTB, but everything has a cost.  If development works on this, something else won't get attention.  Perhaps if I build it, I can send it to them and they can use it.  :-)

  • Well I'm not trying to complain too loud - After all they do provide us a way to work around these sorts of issues.

     

    Good luck