[
https://issues.apache.org/jira/browse/PHOENIX-3860?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16055391#comment-16055391
]
Ohad Shacham commented on PHOENIX-3860:
---------------------------------------
Thanks [[email protected]].
Regarding the secondary index. As far as I understood, while creating an index
to a non empty data table you update the index with the table data (original
data) while using the timestamp taken at the fence and not using the original
data timestamp. Entries of transaction that started after the fence are added
to the index with the same timestamp as appear in the data table.
I assume that the timestamp of the fence should be generated during the fence.
However, taking a look at the code I see that the fence creation does not
return a timestamp and a transaction is started right after creating the fence.
Does the timestamp of this transaction is being used for the index entries of
the original data? If so, can't we violate snapshot isolation if another
transaction started right in between? this transaction timestamp is smaller
than the index entries of the original data even though they were
added/committed after the original data.
Thx,
Ohad
> Implement TAL functionality for Omid
> ------------------------------------
>
> Key: PHOENIX-3860
> URL: https://issues.apache.org/jira/browse/PHOENIX-3860
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Ohad Shacham
> Assignee: Ohad Shacham
>
> Implement TAL functionality for Omid in order to be able to use Omid as
> Phoenix's transaction processing engine.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
