[jira] [Commented] (CASSANDRA-8224) Checksum Gossip state

["Brandon Williams (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/CASSANDRA-8224?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14314229#comment-14314229 ] 

Brandon Williams commented on CASSANDRA-8224:
---------------------------------------------

This is a prime example of why you should have ECC memory, so we don't have to exert ourselves in software to accomplish what hardware can already provide.  That said, the way I see this working is to provide the checksum as a gossip state itself (that way older nodes can just ignore it) which is a checksum of everything except the checksum state itself.  But again, it does feel like a problem better solved elsewhere.

> Checksum Gossip state
> ---------------------
>
>                 Key: CASSANDRA-8224
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-8224
>             Project: Cassandra
>          Issue Type: Improvement
>            Reporter: sankalp kohli
>            Assignee: sankalp kohli
>            Priority: Minor
>
>  We have seen that a single machine with bad memory can corrupt the gossip of other nodes and cause entire cluster to be affected. If we store and pass the checksum of the entire state, we can detect corruption. If a bad machine tries to bump the generation number or other things, it will be detected and ignored.



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