cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Cristian Opris (JIRA)" <>
Subject [jira] [Commented] (CASSANDRA-5062) Support CAS
Date Tue, 05 Mar 2013 17:58:16 GMT


Cristian Opris commented on CASSANDRA-5062:

The read path is not detailed in your document. Also even with 500 lines of code and explanations,
there can be ambiguities.

It's not obvious to me that you intended read to repair N rounds of paxos before responding.
Yes, could work but I don't feel it's 
the best way to do it.

Sorry if you feel that I'm spamming this Jira but:

- It's the only forum I have available. I can't post to the mailing list for objective reasons.
- you guys can be very selective in what comments you actually care to respond to. I've been
going on about version counters and separating rounds for a long time here without getting
any interest but then you come back and propose something very similar. Since you posted it
I thought you're looking for some actual review and feedback.

bq. in practice, learn messages won't be lost randomly as in your example

This is really not an assumption I would make. Why even bother with Paxos if nothing can go
wrong anyway ? 

> Support CAS
> -----------
>                 Key: CASSANDRA-5062
>                 URL:
>             Project: Cassandra
>          Issue Type: New Feature
>          Components: API, Core
>            Reporter: Jonathan Ellis
>             Fix For: 2.0
>         Attachments: half-baked commit 1.jpg, half-baked commit 2.jpg, half-baked commit
> "Strong" consistency is not enough to prevent race conditions.  The classic example is
user account creation: we want to ensure usernames are unique, so we only want to signal account
creation success if nobody else has created the account yet.  But naive read-then-write allows
clients to race and both think they have a green light to create.

This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see:

View raw message