ctubbsii commented on PR #5988:
URL: https://github.com/apache/accumulo/pull/5988#issuecomment-3573393140

   > Edit: It's not `sync on file close`, it's `sync on block close`.
   
   Thank you for the clarification.
   
   I think I'm still confused about the difference between this create mode 
flag in this PR, and the `dfs.datanode.synconclose` property in the datanode 
config. The latter seems obvious: when writing any blocks locally, call sync on 
the local filesystem on closing them (perhaps that's not what it does, but 
seems like it). That seems like it would apply to blocks created for 
replication or for EC blocks... though I suppose the EC code path could just do 
something entirely different. But it's not clear to me what this create mode 
flag is doing inside the HDFS client library. Is that equivalent to 
`dfs.datanode.synconclose`, but just passed along the RPCs to the datanode for 
writing blocks for that particular file, rather than a global config? Or does 
it trigger an extra RPC sync message in the HDFS client library when talking to 
a datanode, like after-the-fact?


-- 
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