If you use clustered cache, the data will be loaded to nodes all over the cluster. That's why when you decide to run the indexing the MR task re-indexes whole cluster. What's the use case for re-indexing only single node?
Or, do you have the data in local caches with the same name and does the re-indexing start on all those (non-related) caches?
Btw., there have been found several problems with MassIndexer after 6.0.2, please check them and possilby upgrade to 7.1.x.
Because we are hitting this issue https://issues.jboss.org/browse/ISPN-4617. We want to upgrade to 7.x but we saw too many api changes and our go live date for our project is <15 days and we don't think we have time to complete the upgrade.
In our use case, we are loading data into the cache (in local mode as we don't want to replicate this data) as part of the service startup and need to index it. so we don't need to reindex on all the nodes in the cluster as each node will do this independently
Thanks in advance
Seems that the implementation was changed from MapReduce-based MassIndexer to DistributedExecutors-based one, so there's no trivial backport. You can try this trivial fix https://github.com/infinispan/infinispan/pull/2786/files , or you can try to backport https://github.com/infinispan/infinispan/pull/3132 . Infinispan 6.x won't have any further bugfix release.