Re:Scaling nodes in a Dgraph cluster
by mrjn on Tuesday June 21, 2016 @09:23AM
(#11945387)
It depends upon the memory available. We have a background goroutine which checks how much memory we're using and evicts posting lists if we exceed a predefined usage. If we have a lot of memory then we just won't evict the posting lists, which means they'll be in RAM.
All this "caching" happens close to the disk level, just above RocksDB. We don't cache any results -- to me caching is cheating -- and we want to build something truly low latency. A user can easily add caching if they want to decrease the load on Dgraph, but it's not something I think we should be doing at the db level.
All this "caching" happens close to the disk level, just above RocksDB. We don't cache any results -- to me caching is cheating -- and we want to build something truly low latency. A user can easily add caching if they want to decrease the load on Dgraph, but it's not something I think we should be doing at the db level.