I have a task to investigate in using cassandra with hadoop.
Hadoop’s namenode is a single point of failure. This means that if the namenode goes down, the entire database is unreachable. All Cassandra nodes are identical so there is no single point of failure.
If I used Cassandra with Hadoop I assume that all the system has a single point of failure (which is the Hadoop namenode as I understand).
My questions:
1- Is my previous assumption true ?
2- If this is the case, is there any advantage for using such structure (Cassandra with Hadoop) ?
Its half-truth. If you are using Cassandra with Hadoop, all your data are stored in Cassandra instead of HDFS and Hadoop is being used for calculation. So if your namenode goes down the only one you lose is ability to run map reduce tasks, but all your data will be safe and accessible.