Portal v7.0.0.2: Content does not replicate to all nodes in a cluster

Content updates in a WebSphere Portal v7.0.0.2 cluster may not replicate to all nodes in a cluster until the Portal JVMs are restarted. This can include, but is not limited to, the following:

• Updates to portlet preferences
• Updates to pages, or new pages created
• Updates to virtual portals, or new virtual portals created

Portal caches this content for each node, and the cached data is not being cleared in a timely manner, so old data is displayed on the impacted nodes. A number of defects have been created to improve this caching, and have been included in cumulative fix 16 (PM67998).

If you experience any issues with content not replicating across cluster members, it is strongly recommended you apply CF 16 (or higher). See 'Recommended Updates for WebSphere Portal' for latest cumulative fix information.

If you still see caching related issues after applying cumulative fix 16 (or higher), then please open a PMR with IBM WebSphere Portal support and be prepared include the following information:

- Current Cumulative Fix level
- Specific use case to recreate the problem. For example you edit links in the Bookmarks portlet and the changes do not replicate to the other JVMs until the JVMs are restarted.

- Any customizations to the CacheManagerService properties

- Any customizations to the WebSphere Dynamic Cache Service settings for each Portal JVM.

- wpcollector collection from each node in the cluster (see link in Related URLs)

- Frequency of issue (does it occur with every content change? Intermittently?)