Flavio Paiva Junqueira commented on ZOOKEEPER-29:

Just to motivate a little more this scheme, taking the example above, suppose 
that each group corresponds to the set of ZooKeeper servers in a colocation. To 
form a quorum, we just need to get a majority of servers from a majority of 
groups. This gives us quorums of size 4, whereas majority quorums would require 
5. This model is reasonable, though, as it is reasonable to expect that at 
least a majority of colocations have a majority of servers up and running.  

This idea actually appeared in the following citation (apologies for the 
self-reference): http://dslab.epfl.ch/hotdep/2005/papers/junqueira_dependent.pdf

Another question that may arise is how leader election changes. With this 
scheme it only changes minimally. In the current trunk code, we always wait for 
a majority of votes, and that's because we have majority quorums hardcoded. 
With this patch, we check if we have a quorum, and a quorum can be a majority, 
as we have currently, or a different quorum system as the one of the example.

> Flexible quorums
> ----------------
>                 Key: ZOOKEEPER-29
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-29
>             Project: Zookeeper
>          Issue Type: New Feature
>          Components: server
>            Reporter: Patrick Hunt
>            Assignee: Flavio Paiva Junqueira
>             Fix For: 3.2.0
>         Attachments: ZOOKEEPER-29.patch, ZOOKEEPER-29.patch, 
> ZOOKEEPER-29.patch, ZOOKEEPER-29.patch, ZOOKEEPER-29.patch, ZOOKEEPER-29.patch
> Moved from SourceForge to Apache.
> http://sourceforge.net/tracker/index.php?func=detail&aid=1938782&group_id=209147&atid=1008547

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

Reply via email to