SA-27038

Summary

In a scenario where one agent performs processing on a work object and queues the same work object for another agent during that processing, broken queue items are sporadically generated with errors indicating that the second agent is unable to obtain a lock on the work object.

Error Messages

Cannot obtain a lock on instance MYCO-ORG-MYAPP-WORK CASE-XXXXX, as Requestor B1234C1A12345678A1A1BF1EA1CA12345 already has the lock

Steps to Reproduce

Unknown

Root Cause

An issue in the custom application code or rules where, after the work object was queued for the second agent by the first agent, the first agent could potentially continue processing the work object for an amount of time longer than the amount of time the second agent waited before attempting to pick up the work object

Resolution

To avoid this issue (reported for PRPC 6.2 SP2), set a longer delay when queuing the work item for the second agent to allow enough time for the first agent to complete its processing and release the lock.

Alternatively, use the Deferred option when queueing the work item for the second agent.
Refer to this PRPC 6.2 Help topic:
https://pdn.pega.com/sites/default/files/help_v62/procomhelpmain.htm#methods/queue-for-agent/queue-for-agent.htm

There might be other problem scenarios (using PRPC 6.2 SP2 or a later product release) similar to the one described here where one approach is preferred over another approach. If you have a similar problem scenario related to agent queues, search the PDN for a solution. If you cannot find a solution for your scenario, post your question to the Product Support Community.

Related Content
https://pdn.pega.com/sites/default/files/help_v73/procomhelpmain.htm#methods/queue-for-agent/queue-for-agent.htm
https://pdn.pega.com/about-agent-queues/about-agent-queues