[
https://issues.apache.org/jira/browse/HADOOP-9421?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13641073#comment-13641073
]
Luke Lu commented on HADOOP-9421:
---------------------------------
bq. RPC connections are already cached and multiplex multiple high level
clients. It's a premature optimization to introduce extra complexity to remove
one RPC for reconnects.
I meant we can even persist the server mechs cache (e.g. in /tmp/hadoop-$user)
for app/job clients which runs in different processes. protobuf makes it
straightforward.
bq. I abandoned it and proposed a hybrid compromise. What are your thoughts on
that?
The "compromise" is definitely an improvement over sasl-next. As long as the
client optimization, which is reasonable IMO, can be added later in a backward
compatible way, I don't feel the need to push for client optimization (to
ensure amortized O(1) round-trip) for this jira :). From impl pov, client
initiation simply always sends the sasl-start request upon connection. If you
want to reduce the impl scope (which is a good thing), you can have server
simply ignore the first sasl-start and always present sasl-mechs.
> Add full length to SASL response to allow non-blocking readers
> --------------------------------------------------------------
>
> Key: HADOOP-9421
> URL: https://issues.apache.org/jira/browse/HADOOP-9421
> Project: Hadoop Common
> Issue Type: Sub-task
> Affects Versions: 2.0.3-alpha
> Reporter: Sanjay Radia
> Assignee: Junping Du
> Attachments: HADOOP-9421.patch
>
>
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
