Question-86: Which all are the filesystem supported by HDFS as of now?

Answer: Following file systems are recommended as of now

• Ext3
• Ext4
• XFS: This is default on the RHEL7
• AWS S3

Question-87: Why it is said that on Linux system file read operations also leads to a write operation?

Answer: Linux filesystem keep metadata that records when each file was accessed. This means that even reads results in a write to the disk. Hence, to speed up the file reads, it is recommended that you disable this feature.

Question-88: Why it is recommended that during the filesystem mount, we should not use the sync option?

Answer: The filesystem mount options have a sync option that allows you to write synchronously. Using the sync filesystem mount option reduces the performance for services that write data to disks, such as HDFS, YARN, Kafka, and Kudu. Because most of the time write operation is already replicated and synchronous writes to disk is unnecessary, expensive, and do not measurably improve stability. 

Question-89: How do you explain the High Availability and Load Balancing?

Answer: Load balancing refers to distributes the operations across multiple services let’s say RDBMS instances in parallel. While HA focuses on the service continuity. However, Load balancing deployment often used as part of HA strategies to overcome demands and monitoring and failover management in HA environment. 

However, in the case of Cloudera it is clearly mentioned that components are not designed to support load balancing. During HA strategy with multiple instances ensure that connections are routed to a single RDBMS service at any given time. 

Question-90: What do you mean by ordinary objects pointers In Java?

Answer: This is an Optimization technique in Java Which is also known as Compressed oops Which enable 64-bit JVM to address heap size up to 32GB using 4-byte pointers.  For large heap size 8-byte pointer are required. This means that heap size slightly less than 32GB can hold more objects than a slightly more than 32GB.

So, it is recommended that if you don't need heap size more than 32GB then use heap size 31GB or less to avoid this issue.  if you need 32GB or more then that your heap size to 48 GB or higher to account for the larger pointers, so whenever you need heap size more than 32GB then multiply the amount of heap you need by 1.5