errose28 commented on PR #3890:
URL: https://github.com/apache/ozone/pull/3890#issuecomment-1297498522
Thanks for the proposal @Cyrill this looks like a good feature to add to
Ozone. Proposal looks good overall I just have a few comments:
- Will we support compressing individual keys as well as buckets? For
example, here is how it works for erasure coding:
- The EC replication type can be set as the default for the bucket.
- The replication type can be overridden at the key level, for example
you can create a Ratis key in an EC bucket and vice versa. If no type is
specified the default is used.
- Supporting client/server cross compatibility
- **old client new server**: Since this is a client side change, we will
need to add a client version (`org.apache.hadoop.ozone.ClientVersion`) so that
a new server does not allow an old client (which does not support compression)
to read compressed data that a new client wrote.
- **new client old server**: It looks like we *might* need a server
version (`org.apache.hadoop.ozone.OzoneManagerVersion`) to handle this case as
well. I assume some metadata is being written on the server side to mark that
the data has been compressed and in what format? If this is stored on the OM
then an old server will not persist this metadata and no clients will be able
to read the compressed data.
--
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]
