[ 
https://issues.apache.org/jira/browse/PHOENIX-2846?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

James Taylor updated PHOENIX-2846:
----------------------------------
    Attachment: PHOENIX-2846.patch

> Fix CreateTableIT.testCreateTableWithoutSchema()
> ------------------------------------------------
>
>                 Key: PHOENIX-2846
>                 URL: https://issues.apache.org/jira/browse/PHOENIX-2846
>             Project: Phoenix
>          Issue Type: Bug
>            Reporter: James Taylor
>            Assignee: Ankit Singhal
>             Fix For: 4.8.0
>
>         Attachments: PHOENIX-2846.patch
>
>
> From your namespace enhancement, you added the 
> CreateTableIT.testCreateTableWithoutSchema() but aren't managing the 
> timestamp yourself as required by tests derived from BaseClientManagedTimeIT 
> through code like this:
> {code}
>         long ts = nextTimestamp();
>         Properties props = new Properties();
>         props.setProperty(PhoenixRuntime.CURRENT_SCN_ATTRIB, 
> Long.toString(ts));
> {code}
> This will lead to other tests failing with a NewTableAlreadyExistsException.
> We're trying to move away from tests that derive from 
> BaseClientManagedTimeIT, so the best fix is to derive from the new 
> BaseHBaseManagedTimeTableReuseIT and follow the same pattern as other tests 
> to get the table name from the test util function. These tests will execute 
> much faster because the underlying HBase metadata never needs to be dropped. 
> You'd need to get rid of the setting of the CURRENT_SCN property and then 
> there's no need to open new connections all the time.
> Another slightly less idea fix is to derive from BaseHBaseManagedTimeIT base 
> class instead, and also get rid of the setting of the CURRENT_SCN property.
> Another fix would be to just set the CURRENT_SCN property for your new test. 
> However, that would not be increasing the greater good as much. :-)
> Appreciate if you could double check for other occurrences of this over your 
> last few check-ins too.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Reply via email to