We are running Ignite 1.7 on 20+ servers and using as distributed cache store. Caches are partitioned without backup (cacheMode=PARTITIONED, backups=0).
>From time to time we are getting network segmentation, so servers are getting stopped as per default segmentation policy. In our case we would like to be able reconnect to cluster instead of stopping server: if we have cache entries with same key on different segments we can use any cache entry or even drop it. Can changing properties like SegmentationResolveAttempts or SegmentCheckFrequency can help? Is it possible to implement custom segmentation policy that will allow reconnect to cluster? And if it is possible can you provide some documentation or direction how to do this. Thank you. -- Yitzhak Molko
