bitterfox commented on a change in pull request #3677:
URL: https://github.com/apache/hadoop/pull/3677#discussion_r757206282
##########
File path:
hadoop-hdfs-project/hadoop-hdfs-client/src/main/proto/datatransfer.proto
##########
@@ -38,6 +38,7 @@ message DataTransferEncryptorMessageProto {
SUCCESS = 0;
ERROR_UNKNOWN_KEY = 1;
ERROR = 2;
+ ERROR_ACCESS_TOKEN = 3;
Review comment:
> If the server sends ERROR_ACCESS_TOKEN and the client is old (i.e.
this patch is not applied), client cannot understand the ERROR_ACCESS_TOKEN and
it is treated as SUCCESS (0),
I see. In such a case, the behavior of the client-side is the same, got some
exception in DFSInputStream and long sleep happens (the exception happens will
be different though). But I agree it's not graceful, I added accessTokenError
bool and set it true with ERROR status. Then old client behaves completely the
same.
> In my understanding, setting SUCCESS as default was a mistake when
creating the enum.
I agree. I changed the handling of status in readSaslMessage to reject
unknown status
--
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]
