Moving stuck IBM Business Process Manager tokens right along
D. Payne 270005S18K Comments (4) Visits (13955)
By David J Payne (IBM Business Process Manager Level 2 Support)
The term “Stuck Token” is used to describe when tokens are on a task that has been completed but is still marked as active. When a token is active and the task where the token is located becomes marked complete, the token should move onto the next task. The status of tokens are found by searching on the LSW_TASK and LSW_BPD_INSTANCE tables for the business process definition (BPD) instance IDs of the tokens in question. Active instances have their EXECUTION_STATUS set as 1 and closed instances have their STATUS set as 32; this is viewed in the LSW_TASK table.
The previous definition and characteristics are important to remember when you monitor the status of running applications because tokens also stop moving for other reasons. Tokens stop moving when application errors occur; with this there are usually error messages that display. Another situation where tokens do not move is when they become "orphaned." This situation typically only happens during the migration of in-flight instances and is easily identified by the token showing up as "Unknown." When process execution halts due to an error in the application and the token is on a failed task, the following solutions will likely not work.
When tokens become stuck during process execution there are several possible methods that will get them moving again. The following possibilities are not guaranteed to resolve stuck tokens on all systems because tokens become stuck for a variety of reasons.
Usually this process is used to push the token forward to the next activity. This process is useful when you are sure that all processing has completed at the task where the stuck token is located. If all processing has not completed, moving a token’s position can possibly cause data loss.
If you have any questions or comments about this blog article, we welcome your feedback below this entry.