Why Batch Processing In The Federation Cluster Takes Longer Than Expected Time? (Doc ID 2572162.1)

Last updated on OCTOBER 13, 2023

Applies to:

Goal

Customer reporter that one of the coherence consuming application has also seen slowness with FederationCacheService compared to regular DistributedCacheService. The consumer has told us that it use to take ~30 minutes to load the batch data to coherence cache. However, since Federation, the same batch took 1hr20 minutes. We have rechecked it by running the batch thrice. My assumptions here are that the FederationCacheService is trying to do two put() operations one to the cache and the other one to 'journal record' queue to facilitate federation to remote side. I would like to know if the journal record is listening to cache events and making its own copy. Alternatively, is that the FederationService doing an explicit write operation to journal record queue as well?

For this particular consumer, we have reverted the cache service to regular 'distributedCacheService' as it used to be and batch load is running (~30min) as normal.

Solution

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.