Summary

When running Confluence Directory synchronization in Data Center, the data center nodes will display the last synchronization from the individual Node, not the cluster-wide last sync time. an Admin looking at only one node may believe the synchronization did not happen, and mistakingly run the manual synchronization process. 

Environment

• Oracle 12c 12.1.0.2.0 - ojdbc7.jar driver
• Confluence 5.10.8 Data Center

Steps to Reproduce

• Connect to LDAP/AD Server
• Allow the system to sync
• Observe the Sync time on Each Node

Expected Result

Each Node should reflect the last time that the directory was synced cluster-wide

Actual Result

Each Node reflects the last time the sync was run on that node. See a2,jpg below:

 * *

And a3,jog below from the same cluster:

 NOTE:   The times are the same, but the dates are different. Reviewing these, an admin might believe that the sync did not run for a day.

Workaround

There is no actual workaround other than a visual inspection of all nodes to verify the last run date of the actual synchronization from the nodes or to run the queries:

SELECT NEXT_RUN_TIME FROM SCHEDULER_CLUSTERED_JOBS AS JOB WHERE JOB.JOB_RUNNER_KEY LIKE %DirectoryPollerManager%;

SELECT * FROM SCHEDULER_CLUSTERED_JOBS AS JOB WHERE JOB.JOB_RUNNER_KEY LIKE %DirectoryPollerManager%;

This will provide the next run time of the sync (which should be in the future) and all of the information about the job.

Resolution

The recommended resolution for customers is to upgrade to the latest version of Confluence or for customers on an LTS release such as 7.4.x, to upgrade to the latest LTS release such as 7.13.x