i have few questions regadring corrupt block ,it would be really great if someone can help me on it.
1)Can anyone let me know what is the specific reason for a corrupt block.
2)If i have a replication factor as 3 , then why do i receive corrupt block alert , doesnt the name node repair the corrupt block by copying a good block from another node?
3)is it okay to remove that particular block/file when i receive this alert ? will the name node then scan and find 2 replicas instead of 3 and copy 1 replica to the node to make it as 3 in order to maintain the replication factor as 3.
A "corrupt" block means some of the block replicas are corrupt. NameNode should be able to schedule replication from healthy replica. The block status report system is sometimes not up to date with the actual block state. (meaning: the block is healthy again but the reporting system like CM may still show bad health) I'd suggest run fsck to make sure.
A missing block means all replicas are corrupt or missing, and you can't recover.
A corrupt file means one of the file's block is missing.