[
https://issues.apache.org/jira/browse/PHOENIX-2605?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15133830#comment-15133830
]
James Taylor commented on PHOENIX-2605:
---------------------------------------
For the regular UPSERT SELECT, we start a transaction for CREATE INDEX. We do a
rollback in MetaDataClient.buildIndex(), but I don't think that's necessary. If
we don't do that, would the same transaction be open and used for the UPSERT
SELECT?
If that's the case, that's probably worth fixing. Also, I think it'd be pretty
easy to do my work around of having a TX_SCN - if nothing else it'd be more
efficient.
Either way, what you have is an improvement, so let's check that in. Give this
some thought and see if you can come up with any scenarios where you'd run into
problems having a timestamp past when the index was created.
> Enhance IndexToolIT to test transactional tables
> ------------------------------------------------
>
> Key: PHOENIX-2605
> URL: https://issues.apache.org/jira/browse/PHOENIX-2605
> Project: Phoenix
> Issue Type: Test
> Reporter: Thomas D'Silva
> Assignee: Thomas D'Silva
> Fix For: 4.7.0
>
> Attachments: PHOENIX-2605.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
