cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Branimir Lambov (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-8268) Token serialization should not assume all tokens are created by the database partitioner
Date Tue, 25 Nov 2014 14:11:13 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-8268?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14224568#comment-14224568
] 

Branimir Lambov commented on CASSANDRA-8268:
--------------------------------------------

Patch uploaded, comparison viewable [here|https://github.com/blambov/cassandra/compare/8268-token-serialization-partitioner].

Moves the selection of partitioner for the de-serialization of {{Token}} and Token-based entities
({{RowPosition}}, {{Bounds}}, {{Range}} etc.) up to the messaging layer, where we can be confident
which partitioner needs to be used. Makes the serialization of {{MerkleTree}} correct in the
case where the partitioner for the tree is different from {{StorageService.getPartitioner()}}.

Removes the last references to {{StorageService.getPartitioner()}} from the Token-based classes.


> Token serialization should not assume all tokens are created by the database partitioner
> ----------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-8268
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8268
>             Project: Cassandra
>          Issue Type: Bug
>            Reporter: Branimir Lambov
>            Assignee: Branimir Lambov
>            Priority: Minor
>         Attachments: 8268.patch
>
>
> This is also a hidden reference to StorageService.getPartitioner within Token, but this
needs to be handled more carefully as we need to check whether tokens serialized with the
wrong partitioner could be present in user data.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message