Currently we are using infinispan 7.0.3.Final version . We have a replicated cache (file-based-persistence) and its clustered. We also use jgroups to communicate between nodes.
We recently saw loss of indices and decided to use MassIndexer to recover the indices. But it does seem to work in a clustered environment and gives us random numbers with loss of data. Only a fresh reload of cache data helps us in recovering the indexes and its a costly operation.
Can anyone suggest how to load the file based indices for a particular node again?
The MassIndexer has been revamped in version 8+ and works in clustered caches with all configurations, apart from being faster. I'd recommend upgrading Infinispan to 8.2 or 9.1.1.Final.