ESP_UPDATE Failed

Aaron Van

Is there any way to pull a little more info out of these failures in the logs?

 

Details:Stored procedure 'esp_update_action' failed.

 

What exactly is the esp_update_aciton supposed to do and when it says it fails does that mean the entire transaction is roled back?

 

thanks,

 

...aaron

Ryan Saul

I get these error very frequenty (serveral times a day).

 

I worked out my esp_update_action was when the actions did not refresh against the stage for a user whilst progressing through a process.. the error is then prompted when hitting an 'obsolete' action.

 

I had queried this previously and logged a SR with no success to resolve.

Jerome_Pearce

We also get these a lot. After a manual data migration to version 9, if actions have been renamed, we get these. This was how we initially found the eWait entries we needed to rename. It also occurs in unmigrated systems where erroneous eWait entries exist (for whatever reason). I think it also happens when chained actions fail because the action name is incorrect.

 

In most cases it appears to be because of an incorrect action name, and these are the only instances I am aware of where this can occur. Having said that, it is difficult to diagnose because often the error log does not contain enough information.

Aaron Van

The errors don't seem be causing any issue with the workflow, but I'd like to know the root cause in case the result ever become a little more violent.

Robert Vignerot

As both Saul and Jerome mention, this occurs when the action the engine is trying to invoke is not available for the current stage.  I think that the most common cause is contention as Saul mentions.   This is usually where 2 users open a folder at the same time.  User 1 completes an action moving the folder to another stage and then the second folder tries to invoke an action that is no longer available.  We were getting litterally thousands of these per day as users tried to open items off of a very high volume shared queue. 

 

This was a significant issue for us and we addressed this by adding a locking mechanism, whereby a user would lock a set of folders via a flag action and then work off the set of folders assigned to them.  There are still some of these errors, but they now occur on the flag action and no longer impact the users directly.