[
https://issues.apache.org/jira/browse/HBASE-18966?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16203309#comment-16203309
]
Chia-Ping Tsai commented on HBASE-18966:
----------------------------------------
bq. is it like that is already happening?
That is already happening...see
[Segment#updateMetaInfo|https://github.com/apache/hbase/blob/d35d8376a70a8de63c5d232a46e39657ba739eef/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/Segment.java#L283].
bq. If so pls change the jira subject/desc accordingly.
done. Thanks for the reminder.
> Use non-sync TimeRangeTracker as a replacement for TimeRange in
> ImmutableSegment
> --------------------------------------------------------------------------------
>
> Key: HBASE-18966
> URL: https://issues.apache.org/jira/browse/HBASE-18966
> Project: HBase
> Issue Type: Sub-task
> Reporter: Chia-Ping Tsai
> Assignee: Chia-Ping Tsai
> Fix For: 2.0.0-beta-1
>
> Attachments: HBASE-18966.v0.patch, HBASE-18966.v1.patch,
> HBASE-18966.v2.patch, HBASE-18966.v2.patch
>
>
> The in-memory compaction/merge do the great job of optimizing the memory
> layout for cells, but they don't update its {{TimeRange}}. It don't cause any
> bugs currently because the {{TimeRange}} is used for store-level ts filter
> only and the default {{TimeRange}} of {{ImmutableSegment}} created by
> in-memory compaction/merge has the maximum ts range.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
