[
https://issues.apache.org/jira/browse/HBASE-26350?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17427365#comment-17427365
]
Josh Elser commented on HBASE-26350:
------------------------------------
I know this looks like a silly/stupid change (how simple it is), but it goes a
long way in figuring out what went wrong server side. I need to look at the raw
SASL reply we do to see if the exception is actually getting propagated back to
the client properly (before the SASL handshake completes).
Because replication was in the mix for me, it's just as tough to get the client
exception as it is the server exception. At least I can make one of those much
easier ;)
> Missing server side debugging on failed SASL handshake
> ------------------------------------------------------
>
> Key: HBASE-26350
> URL: https://issues.apache.org/jira/browse/HBASE-26350
> Project: HBase
> Issue Type: Bug
> Reporter: Josh Elser
> Assignee: Josh Elser
> Priority: Minor
> Fix For: 2.5.0, 3.0.0-alpha-2, 2.3.7, 2.4.8
>
>
> In trying to debug some problems with the pluggable authentication, I noticed
> that we are eating the IOException without logging it (at any level) in
> ServerRpcConnection.
> This makes it super hard to debug when that pluggable interface has a problem
> because the context gets lost (clients just get a pretty useless DNRIOE).
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
