[
https://issues.apache.org/jira/browse/PHOENIX-2005?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14557508#comment-14557508
]
Gabriel Reid commented on PHOENIX-2005:
---------------------------------------
Took a look through the patch, looks good to me in general, although I haven't
specifically tested any of this stuff.
It probably wouldn't be bad to add some more testing of having a custom znode
and custom port -- it looks like there are a few tests covering this, but not
many (or maybe I just didn't see them).
Lots of nice little typo/spelling fixes in there too.
One thing that I noticed is that the CsvBulkLoadToolIT (and possibly other
tests) does stuff like this:
{code}csvBulkLoadTool.setConf(new
Configuration(hbaseTestUtil.getConfiguration()));{code}
which means that it injects things like {{hbase.zookeeper.property.clientPort}}
into the underlying configuration, which will probably circumvent some testing
of correctly interpreting alternate ports. Ideally we could run some tests (or
at least the CsvBulkLoadToolIT) with an alternate hbase znode, although I don't
know if that's possible via the HBaseTestUtil.
> Connection utilities omit zk client port, parent znode
> ------------------------------------------------------
>
> Key: PHOENIX-2005
> URL: https://issues.apache.org/jira/browse/PHOENIX-2005
> Project: Phoenix
> Issue Type: Bug
> Reporter: Nick Dimiduk
> Assignee: Nick Dimiduk
> Fix For: 5.0.0, 4.5.0
>
> Attachments: PHOENIX-2005.00.patch
>
>
> Our config parsing utilities assume the zookeeper quorum server list is
> sufficient for connecting to HBase. This not the case and should not be
> assumed; take into account client port and znode parent. See conversation
> over on PHOENIX-1980.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
