I think the client assumes the data is written after $dfs.namenode.replication.m...

growse on May 2, 2015 | parent | context | favorite | on: Elasticsearch node crashes can cause data loss

I think the client assumes the data is written after $dfs.namenode.replication.min blocks have been written, which I think is 1 by default.

What it actually means inside HDFS when it claims 'written', I'm not sure - I'd assume flushed to the dirty page buffer at a minimum and would hope fsync.