[
https://issues.apache.org/jira/browse/HBASE-15389?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15184570#comment-15184570
]
Clara Xiong commented on HBASE-15389:
-------------------------------------
1. We currently use DTCP with DefaultCompactor, which only output a single file
from the file list DTCP picks, so that we can compact within every date-tiered
window into a single file and the resulted layout will be tiered by date.
We will use this new compactor WITH DTCP to output date tiered layout if the
file(s) to compact expands over multiple time windows such as a major
compaction or bulk load files.
So DTCP is not going away. By pairing it with this new compactor, we can
improve for some use cases. HBASE-15400 has more details.
2. The windowing logic in DateTieredMultiFileWriter you referred to is to
output data to multiple files based on a given list of boundaries. These
boundaries will be determined in DTCP and passed as part of
DateTieredCompactionRequest it generates to DateTieredCompactor.
> Write out multiple files when compaction
> ----------------------------------------
>
> Key: HBASE-15389
> URL: https://issues.apache.org/jira/browse/HBASE-15389
> Project: HBase
> Issue Type: Sub-task
> Components: Compaction
> Affects Versions: 2.0.0, 1.3.0, 0.98.19
> Reporter: Duo Zhang
> Assignee: Duo Zhang
> Fix For: 2.0.0, 1.3.0, 0.98.19
>
> Attachments: HBASE-15389-uc.patch, HBASE-15389-v1.patch,
> HBASE-15389-v2.patch, HBASE-15389.patch
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
