[
https://issues.apache.org/jira/browse/HBASE-12859?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14296211#comment-14296211
]
Lars Hofhansl commented on HBASE-12859:
---------------------------------------
Thanks [~stack].
bq. nit: getLastMajorCompactionTsForTable should have a comment explaining it
can return zero and what that means.
Yes, I need to add that. Thanks for the reminder!
bq. I don't get this bit (or in other words, please explain it to me):
bq. How is ts of oldest HFile the last major compaction time?
I need to change the wording in the comment.
What we're finding is the creation time of the oldest HFile that resulted from
a major compaction. That creation time will the a time before which we are
guaranteed that all data was major compacted. If we do not find an HFile
resulting from a major compaction we have no information and return 0 (i.e. the
beginning of time, maybe I should return HConstants.OLDEST_TIMESTAMP instead)
bq. This attribute of crate time will only be available in certain hfile
versions?
Not completely sure what you mean. It just in the FileInfo. So if written we
have the data, if not we have no information.
Lemme look through the code to make sure it deals fine with that data not being
available.
bq. I'd think the fileinfo already loaded so this is a noop?
It's the only way to get a reference to the fileInfo, if already loaded it'll
used the cached version.
I'll make a new patch soon - will also look at the new findbugs warning.
> Major compaction completion tracker
> -----------------------------------
>
> Key: HBASE-12859
> URL: https://issues.apache.org/jira/browse/HBASE-12859
> Project: HBase
> Issue Type: Brainstorming
> Reporter: Lars Hofhansl
> Assignee: Lars Hofhansl
> Fix For: 2.0.0, 0.98.10, 1.1.0
>
> Attachments: 12859-v1.txt, 12859-v2.txt, 12859-v3.txt, 12859-v4.txt,
> 12859-v5.txt, 12859-v6.txt, 12859-wip-UNFINISHED.txt
>
>
> In various scenarios it is helpful to know a guaranteed timestamp up to which
> all data in a table was major compacted.
> We can do that keeping a major compaction timestamp in META.
> A client then can iterate all region of a table and find a definite
> timestamp, which is the oldest compaction timestamp of any of the regions.
> [~apurtell], [~ghelmling], [~giacomotaylor].
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
