Hi, I had a couple of questions about the HDFS architecture:
1. How does HDFS handle appending to files ? What if there are multiple clients (let us say in the range of 100-200 clients) trying to concurrently append to a file in HDFS. Should this be a performance concern ? Does access need to serialize at the application level, or HDFS knows how to handle requests like these ? 2. Also, I have read about how HDFS is "rack-aware". What if HDFS was deployed to an environment where the physical machines are not local to a rack but exist on a network somewhere ? Is this also a concern ? Thank you.
