bbeaudreault commented on PR #4610:
URL: https://github.com/apache/hbase/pull/4610#issuecomment-1183638114

   @apurtell this is ready for review now. The last build was successful, I 
just pushed a small cleanup.
   
   @Apache9 you may be interested in this. I know you aren't super familiar 
with the BucketCache, but this is mostly just related to how we read 
HFileBlocks from disks.
   
   I think there is another potential solution here which is likely a lot 
simpler -- Similar to https://github.com/apache/hbase/pull/4453/, modify 
HFileBlockDefaultDecodingContext.prepareDecoding to always fill out the 
checksum bytes. In that PR he fills out the bytes with zero's. It seems like if 
we're going to leave space for the checksum, we should instead just do another 
`BlockIOUtils.readFullyWithHeapBuffer` for the remaining checksum portion of 
the `onDiskBlock` ByteBuff so that the output buffer is at least accurate.
   
   That said, this solution is more cpu and space efficient, because we aren't 
leaving space for or copying over a checksum that we will never use.


-- 
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]

Reply via email to