[
https://issues.apache.org/jira/browse/ARTEMIS-209?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15995190#comment-15995190
]
Robbie Gemmell commented on ARTEMIS-209:
----------------------------------------
Yep. Related to https://issues.apache.org/jira/browse/PROTON-900. Most folks
seemed to think it was fine to do that, since it shouldnt make any difference
to the other side (fail is fail..), but I dislike it and raised it anyway. I
later came to realise it was semi deliberate to allow for other behaviours,
though perhaps we can find a middle ground for making it less silly the rest of
the time.
> [AMQP] Broker sends frames after SASL failure
> ---------------------------------------------
>
> Key: ARTEMIS-209
> URL: https://issues.apache.org/jira/browse/ARTEMIS-209
> Project: ActiveMQ Artemis
> Issue Type: Improvement
> Components: AMQP
> Affects Versions: 1.0.0
> Environment: SASL negotiation over AMQP
> Reporter: Chuck Rolke
> Assignee: Timothy Bish
>
> The client sends bogus credentials to the Artemis server. The server
> correctly fails with sasl.outcome code: auth(1). So far so good. Then the
> server sends an AMQP protocol negotiation frame as if everything was OK.
> After failing SASL the server should close the connection.
> Trace file:
> https://people.apache.org/~chug/artemis/20150821-1/artemis-sasl-fail-but-sends-amqp-header.html
> From the trace:
> {noformat}
> 10.10.1.1:1340 -> 10.10.10.254:5672 -> init SASL (3): (1.0.0)
> 10.10.1.1:1340 -> 10.10.10.254:5672 -> method sasl.init
> 10.10.1.1:1340 <- 10.10.10.254:5672 <- init SASL (3): (1.0.0), method
> sasl.mechanisms, method sasl.outcome
> 10.10.1.1:1340 <- 10.10.10.254:5672 <- init AMQP (0): (1.0.0)
> {noformat}
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
