[
https://issues.apache.org/jira/browse/PHOENIX-1390?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14188036#comment-14188036
]
James Taylor commented on PHOENIX-1390:
---------------------------------------
bq. msSinceLastUpdate = rs.getLong(1) - rs.getLong(2);
I'm subtracting two millisecond values. The query changed as well. I realized
later that the subtraction I was doing in the query would give me a DECIMAL
value in units of DAYS. Rather than convert, this was easier.
bq. And in MetaDataEndPointImpl this is removed - incrementTableTimestamp().
This was renamed to clearTableFromCache(). It doesn't need to increment the
table timestamp any longer. Instead, the timestamp of the table takes into
account the timestamp of the stats. Otherwise, when background tasks occur such
as compaction and splits, the client wouldn't get updated stats (since it would
look like the timestamp of the table wasn't different). For these background
tasks, we can't increment the table timestamp. This is a better solutions IMHO.
> Stats not updated on client after major compaction
> --------------------------------------------------
>
> Key: PHOENIX-1390
> URL: https://issues.apache.org/jira/browse/PHOENIX-1390
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
> Assignee: James Taylor
> Fix For: 5.0.0, 4.2, 3.2
>
> Attachments: PHOENIX-1390.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
