[
https://issues.apache.org/jira/browse/HBASE-17291?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15769755#comment-15769755
]
Anastasia Braginsky commented on HBASE-17291:
---------------------------------------------
Hey, [~ram_krish],
I have looked on the RB, generally looks fine. I have two following questions:
1. I do not understand where the close() of the SnapshotScanner should happen
in the common path. I see that in the tests you call for its close() method
explicitly. Where the close() method is invoked when not in the test?
2. I assume you will need to wait till HBASE-17081 is finally committed (after
it was reverted) till you can commit this one, right?
Thanks,
Anastasia
> Remove ImmutableSegment#getKeyValueScanner
> ------------------------------------------
>
> Key: HBASE-17291
> URL: https://issues.apache.org/jira/browse/HBASE-17291
> Project: HBase
> Issue Type: Improvement
> Components: Scanners
> Affects Versions: 2.0.0
> Reporter: ramkrishna.s.vasudevan
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0
>
> Attachments: HBASE-17291.patch, HBASE-17291_1.patch,
> HBASE-17291_2.patch
>
>
> This is based on a discussion over [~anastas]'s patch. The MemstoreSnapshot
> uses a KeyValueScanner which actually seems redundant considering we already
> have a SegmentScanner. The idea is that the snapshot scanner should be a
> simple iterator type of scanner but it lacks the capability to do the
> reference counting on that segment that is now used in snapshot. With
> snapshot having mulitple segments in the latest impl it is better we hold on
> to the segment by doing ref counting.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
