[
https://issues.apache.org/jira/browse/HBASE-30159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18083064#comment-18083064
]
JinHyuk Kim commented on HBASE-30159:
-------------------------------------
[~meszibalu] Hi, I recently worked on adding the XXH3 hash as well, so this
ticket looked interesting to me.
If you're not already working on it, would you mind if I give it a try?
> Make hash algorithm configurable for HashTable/SyncTable
> --------------------------------------------------------
>
> Key: HBASE-30159
> URL: https://issues.apache.org/jira/browse/HBASE-30159
> Project: HBase
> Issue Type: Sub-task
> Reporter: Balazs Meszaros
> Priority: Major
>
> The HashTable MapReduce job utilizes MD5 hashes to determine whether row data
> needs to be synchronized between source and target tables. However, because
> users can control all hash inputs (row, column family, column, value, and
> timestamp), a malicious user could intentionally trigger a collision by
> creating two distinct rows with identical hashes. This would result in less
> rows in the target table which is hard to find.
> We must make the hashing algorithm configurable to allow for more
> collision-resistant hash algorithms.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
