[
https://issues.apache.org/jira/browse/PHOENIX-4449?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16288267#comment-16288267
]
Josh Elser commented on PHOENIX-4449:
-------------------------------------
bq. Do we really all this stuff with try/import? Copying argparse to bin
directory works just fine for me. Have I missed something?
The goal is that we use the system-installed argparse when it is available.
This would avoid us accidentally using "our" copy of argparse when PYTHONPATH
contains {{"."}}
bq. We should not provide any default value for the quorum in the python code.
Otherwise, values from hbase-site.xml will be ignored all the time. What is the
reason for execution an additional process to obtain the quorum address if it
will be done automatically in phoenix.
Ok, that's my confusion. I expected not providing a URL on the command line
would result in a malformed JDBC url being used. You're saying that it the
PhoenixDriver will automatically fill these in?
> Incorrect behavior of sqlline after PHOENIX-3567
> ------------------------------------------------
>
> Key: PHOENIX-4449
> URL: https://issues.apache.org/jira/browse/PHOENIX-4449
> Project: Phoenix
> Issue Type: Bug
> Affects Versions: 4.10.0, 4.13.0, 5.0.0
> Reporter: Sergey Soldatov
> Assignee: Josh Elser
> Priority: Critical
> Fix For: 5.0.0, 4.14.0
>
> Attachments: PHOENIX-4449.001.patch
>
>
> In PHOENIX-3567 we have introduced a usage of argparse in sqlline as well as
> some default values for the parameters. That's not good because of:
> 1. argparse is not a default package in old versions of python like 2.6.6
> which is using in centos 6 distribution.
> 2. We should not set a default value for zookeeper parent because if it's not
> specified in the command line, those values should be obtained from hbase
> client (hbase-site.xml or default values if it's not in the classpath).
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
