cassandra-commits mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Benedict (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (CASSANDRA-7032) Improve vnode allocation
Date Wed, 10 Dec 2014 16:41:13 GMT

    [ https://issues.apache.org/jira/browse/CASSANDRA-7032?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14241349#comment-14241349
] 

Benedict commented on CASSANDRA-7032:
-------------------------------------

If you mean for V vnode tokens in ascending order [0..V), and e.g. D disks, the disks would
own one of the token lists in the set { [dV/D..(d+1)V/D) : 0 <= d < D }, and you guarantee
that the owned range of each list is balanced with the other lists, this seems pretty analogous
to the approach I was describing and perfectly reasonable.

The main goal is only that once a range or set of vnode tokens has been assigned to a given
resource (disk, cpu, node, rack, whatever) that resource never needs to reassign its tokens.

> Improve vnode allocation
> ------------------------
>
>                 Key: CASSANDRA-7032
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-7032
>             Project: Cassandra
>          Issue Type: Improvement
>          Components: Core
>            Reporter: Benedict
>            Assignee: Branimir Lambov
>              Labels: performance, vnodes
>             Fix For: 3.0
>
>         Attachments: TestVNodeAllocation.java, 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
(v6.3.4#6332)

Mime
View raw message