hey all, I want to run a experimental cluster but my machines have limited disk space capacity. I want each node in my cluster to have around 50,000 thousand blocks.
I don't want to have smaller the block size (1K, 4K, etc). I saw SimulatedFSDataset in HDFS code base. Could anybody shed some light in how to use this in a real cluster, i.e. a cluster with everything the same but simulated block? any hint is appreciated. thanks a lot. Thanh
