[
https://issues.apache.org/jira/browse/CASSANDRA-807?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12836313#action_12836313
]
Eric Evans commented on CASSANDRA-807:
--------------------------------------
You've added a call to get_string_property() that is otherwise not needed in
order to force connection errors to happen sooner. I wonder if it wouldn't be
better to just stop using the misleading term "connect(ed)" instead? There is
also an impending change in the pipeline that will make the login() method
required, I'd also be open to implementing that behavior in the cli now as a
solution to this.
Also, Cassandra's coding conventions put opening braces on a line of their own
and this patch moves a number of them up one line. Please see the
http://wiki.apache.org/cassandra/CodeStyle.
> Improve cassandra-cli error display and thrift connection checking
> ------------------------------------------------------------------
>
> Key: CASSANDRA-807
> URL: https://issues.apache.org/jira/browse/CASSANDRA-807
> Project: Cassandra
> Issue Type: Improvement
> Components: Tools
> Environment: Java
> Reporter: Mark Wolfe
> Attachments: cassandra-cli-cleanup.patch
>
>
> A new user observed that thier was some ambiguous error messages presented by
> the cassandra CLI if you connected to the wrong port on the cassandra daemon,
> or in fact anything else listening for a tcp connection. The CLI will say
> "connected" but when you run any of the commands a thrift exception will be
> displayed.
> So I suggest the following enhancements:
> * Verify that the CLI is indeed connected to a thrift port by requesting the
> cluster name
> * Hide as many of the stack traces as possible displaying a simple error
> message
> * Implement a --debug argument to the CLI which will output stacktraces
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
