[
https://issues.apache.org/jira/browse/HBASE-15554?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
ramkrishna.s.vasudevan updated HBASE-15554:
-------------------------------------------
Attachment: HBASE-15554_9.patch
Updated patch. All the duplicates have been removed in the Hash.java class. A
generic HashByteExtractor is used whose subclasses can work with byte[], BB and
cell. Currently the ByteArrayHashByteExtractor and CellBasedExtractor are used.
Once this is committed can do the changes for the Read path such a way that the
cell can be used directly so that there is no byte[] version.
The suggested Iterable way I found it is not suitable here because the Hash
algo knows which byte to be extracted from which offset. The offset calculation
the algo does. We only need to ensure we fetch that byte from that offset from
the underlying structure.
CellUtil has getRowOffset, getQualOffset and getFamOffset but all are marked
Private. Could not find a suitable place to move them as there is no Util for
Bloom in hbase-common.
> StoreFile$Writer.appendGeneralBloomFilter generates extra KV
> ------------------------------------------------------------
>
> Key: HBASE-15554
> URL: https://issues.apache.org/jira/browse/HBASE-15554
> Project: HBase
> Issue Type: Sub-task
> Components: Performance
> Reporter: Vladimir Rodionov
> Assignee: ramkrishna.s.vasudevan
> Fix For: 2.0.0
>
> Attachments: HBASE-15554.patch, HBASE-15554_3.patch,
> HBASE-15554_4.patch, HBASE-15554_6.patch, HBASE-15554_7.patch,
> HBASE-15554_9.patch
>
>
> Accounts for 10% memory allocation in compaction thread when BloomFilterType
> is ROWCOL.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
