[
https://issues.apache.org/jira/browse/CASSANDRA-7032?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Benedict updated CASSANDRA-7032:
--------------------------------
Attachment: TestVNodeAllocation.java
Another updated version with improved statistics on disk distribution. Not up
to scratch for CASSANDRA-6696 unfortunately.
However there's a simpler algorithm still that's much better and trivial to
implement: when we randomise our token selection, instead of randomising each
token over the global token range, simply split the token range into vnode
regions, and randomise a token uniformly within each region. This actually
gives near optimal behaviour across all cluster sizes - disk distributions are
still not perfect, on average having as much as 30% gap in max-to-min range
allocated to each disk with many disks, but this can be improved by having more
vnodes.
> Improve vnode allocation
> ------------------------
>
> Key: CASSANDRA-7032
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7032
> Project: Cassandra
> Issue Type: Improvement
> Components: Core
> Reporter: Benedict
> Labels: performance, vnodes
> Fix For: 3.0
>
> Attachments: TestVNodeAllocation.java, TestVNodeAllocation.java
>
>
> It's been known for a little while that random vnode allocation causes
> hotspots of ownership. It should be possible to improve dramatically on this
> with deterministic allocation. I have quickly thrown together a simple greedy
> algorithm that allocates vnodes efficiently, and will repair hotspots in a
> randomly allocated cluster gradually as more nodes are added, and also
> ensures that token ranges are fairly evenly spread between nodes (somewhat
> tunably so). The allocation still permits slight discrepancies in ownership,
> but it is bound by the inverse of the size of the cluster (as opposed to
> random allocation, which strangely gets worse as the cluster size increases).
> I'm sure there is a decent dynamic programming solution to this that would be
> even better.
> If on joining the ring a new node were to CAS a shared table where a
> canonical allocation of token ranges lives after running this (or a similar)
> algorithm, we could then get guaranteed bounds on the ownership distribution
> in a cluster. This will also help for CASSANDRA-6696.
--
This message was sent by Atlassian JIRA
(v6.2#6252)
