What does fsck do when we run it in HDFS?

Question

I understand what 'fsck' does in local filesystems. How is that logic implemented in Hadoop?

Answer 1

In Hadoop, when we use the 'fsck' command inside the HDFS, the inconsistencies present inside the HDFS data are checked for and if there are certain issues being encountered, the problems (eg: underreplicated blocks or missing blocks) are detected and reported. 

If the failure is recoverable, the NameNode will resolve it after the issue has been reported by the tool.