tomscut commented on pull request #3928:
URL: https://github.com/apache/hadoop/pull/3928#issuecomment-1022794357


   Thanks @sodonnel for your comments and detailed explanations.
   
   I'm sorry I said the wrong number 5M, we have more than 10 million blocks 
per node, and 20 disks, so more than 500,000 blocks per storage.
   
   Your explanation makes sense to me. During `scanDatanodeStorage`, there is 
`listRemove` behavior at the same time, so there will indeed be inconsistency.
   
   Can we copy the blocks of each storage to a `list` first? This is a 
lightweight operation. Then scan the `list` and release the lock once for part 
of scan to reduce the lock holding time. After scan, clear the `list`. Do you 
think this is feasible?


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to