This issue only occurs on an application server cluster deployment of RSA Identity Governance & Lifecycle.
After making a modification to an RSA Identity Governance & Lifecycle feature that uses the scheduler, the item may be duplicated in the in-memory scheduler. This may cause multiple (duplicate) runs of the scheduled task. The duplicate scheduled item may occur at the same time or at different times. If multiple edits are made multiple duplicates can be created.
This issue may affect any item that is scheduled including but not limited to Rules, Collectors (Identity Collectors, Account Collectors, Entitlement Collectors, etc), and tasks. The symptoms may be different depending on the object. There are no error messages related to the duplicate runs although the logs may show duplicate runs where only a single run was expected.
The duplicate only exists in the in-memory copy of the schedule, and the issue can be observed on the UI > Admin > Monitoring > Schedule Information > "Scheduled tasks in memory for "system node only" scheduler" section.
This issue occurs on a cluster deployment when making a change from the non-SON node (General node) to any item that updates the scheduler (even if there is no change to the schedule).
An update action on an existing scheduled item first creates a new item (for the new time) in the scheduler and then deletes the old value. This works correctly when initiated from the SON node. If the request is made from the non-SON node (General node), the database update is completed but the SON node does not correctly delete the old scheduled item.
This is a known issue in the following versions:
RSA Identity Governance & Lifecycle 7.2.0
RSA Identity Governance & Lifecycle 7.2.1
RSA Identity Governance & Lifecycle 7.5.0
This issue is resolved in the following versions which corrects the issue with the scheduler updates:
RSA Identity Governance & Lifecycle 7.2.0 P10
RSA Identity Governance & Lifecycle 7.2.1 P06
RSA Identity Governance & Lifecycle 7.5.0 P02
This issue only affects the scheduled tasks in memory for "system node only" scheduler. The database has the correct information.
Restarting the SON node will remove any existing duplicates from the schedule.
This issue may be avoided by making edits, to items that are scheduled, from ONLY the SON node.
Note that the RSA Identity Governance & Lifecycle scheduler only runs on the SON node in the cluster. Changes made on the non-SON node (General node) to a scheduled item are communicated to the SON node.