sodonnel commented on PR #7009:
URL: https://github.com/apache/ozone/pull/7009#issuecomment-2273650123

   > This section of code selects the minimum BlockGroupLength, thus choosing 
the BlockGroupLength of DataBlock3, leading to the recovery of an incorrect 
data block.
   
   I am not sure this statement is correct.
   
   When a stripe is written, the client must wait for all DNs to ack the write, 
indicating it was successfully saved. If any of the DNs do not return 
successfully, the stripe is abandoned by the client, and a new block is 
requested and the stripe is written to the new block again. This may duplicate 
some data.
   
   To know for sure if this is happening, you need to look at the block length 
stored in OM for this block and see if it aligns with 2 strips (as stated by DN 
3) or greater than 3 stripes, indicating that the "failed write handling" on 
the client did not do the correct thing. Unfortunately given the block ID, its 
not easy to find the key is it associated with. I believe it can be done via 
recon.


-- 
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