[ 
https://issues.apache.org/jira/browse/CASSANDRA-17164?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17482995#comment-17482995
 ] 

Benedict Elliott Smith commented on CASSANDRA-17164:
----------------------------------------------------

1) To persist to the {{promised}} (actually {{promisedWrite}}) register safely 
we would need to modify the timestamp behaviour, i.e. writing promises with a 
lower timestamp than accepts or commits, which harms seamless downgrades and 
backwards compatibility. Not doing so safely obviates any benefit, as we still 
need to perform this comparison in memory to resolve the equal timestamp case.
2) Separate {{accept}} and {{commit}} registers are used for essentially the 
same reason - I agree it would be simpler to maintain just one register here 
for both the mutation and the ballot, but the current representation is 100% 
downgradeable, and any modification to persistence would not be.

We _could_ front-load these changes to the loading from disk phase, and modify 
the {{Snapshot}}, however, if that works for you? This could also modify the 
network representation. I think these changes would be fine, although I think 
they result in only a modest clarity improvement.

> CEP-14: Paxos Improvements
> --------------------------
>
>                 Key: CASSANDRA-17164
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-17164
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Consistency/Coordination, Consistency/Repair
>            Reporter: Benedict Elliott Smith
>            Assignee: Benedict Elliott Smith
>            Priority: Normal
>             Fix For: 4.1
>
>
> This ticket encompasses work for [CEP-14|
> https://cwiki.apache.org/confluence/display/CASSANDRA/CEP-14%3A+Paxos+Improvements].



--
This message was sent by Atlassian Jira
(v8.20.1#820001)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to