You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Darvin Denmian <da...@gmail.com> on 2012/10/01 14:56:36 UTC
Re: Rebalancing cluster
Hi,
1) Yes. I'm using Random Partitioner
2) All 3 nodes in the cluster have used "auto_bootstrap" to acquire their
part of data (I really don't know if this is correct).
Regards.
On Sat, Sep 29, 2012 at 5:56 AM, Bradford Toney <br...@gmail.com>wrote:
> Are you using the random partitoner? Did you set up the tokens the way
> the manual says?
>
> Sent from my iPhone
>
> On Sep 28, 2012, at 7:55 AM, Darvin Denmian <da...@gmail.com>
> wrote:
>
> > Hi,
> >
> > currently my Cassandra cluster has 3 nodes and I'm facing a very common
> problem... Rebalance the data of the cluster.
> > At this point I have:
> >
> > node A: 37,69% of data
> > node B: 7,95% of data
> > node C: 54,36% of data
> >
> > all of them running Cassandra 1.1.3.
> >
> > Can somebody suggest something to solve this problem?
> >
> > Regards.
>
Re: Rebalancing cluster
Posted by Darvin Denmian <da...@gmail.com>.
Hey,
now I got it. I'll try it this night :)
Thanks for all replies.
Regards.
On Mon, Oct 1, 2012 at 11:08 AM, Hiller, Dean <De...@nrel.gov> wrote:
> Nodetool has a move command so you can move to a new better token. Read
> up on the documentation there. I have not used it yet myselfÅ .good idea
> to test it on your test cluster first.
>
> Dean
>
> On 10/1/12 8:03 AM, "Darvin Denmian" <da...@gmail.com> wrote:
>
>>as you can see there is no "Zero Token". Maybe I did something wrong.
>>Will I need to "reset" my entire cluster?
>
Re: Rebalancing cluster
Posted by "Hiller, Dean" <De...@nrel.gov>.
Nodetool has a move command so you can move to a new better token. Read
up on the documentation there. I have not used it yet myselfÅ .good idea
to test it on your test cluster first.
Dean
On 10/1/12 8:03 AM, "Darvin Denmian" <da...@gmail.com> wrote:
>as you can see there is no "Zero Token". Maybe I did something wrong.
>Will I need to "reset" my entire cluster?
Re: Rebalancing cluster
Posted by Darvin Denmian <da...@gmail.com>.
When I execute the command "nodetool ring" the following output appears:
Address DC Rack Status State Load
Effective-Ownership Token
159774252176126302045611947652425555499
10.36.214.118 datacenter1 rack1 Up Normal 16.62 GB
54.36% 82119805777845627277766502152998847873
10.36.214.66 datacenter1 rack1 Up Normal 6.44 GB
7.95% 95648023217342394827789894311936282285
10.36.214.122 datacenter1 rack1 Up Normal 10.85 GB
37.69% 159774252176126302045611947652425555499
as you can see there is no "Zero Token". Maybe I did something wrong.
Will I need to "reset" my entire cluster?
Regards.
On Mon, Oct 1, 2012 at 10:18 AM, Hiller, Dean <De...@nrel.gov> wrote:
>
> You should check the cassandra.yaml file. There is an initial_token in that file that you should have set. The comment above that property reads
>
> # You should always specify InitialToken when setting up a production
> # cluster for the first time, and often when adding capacity later.
> # The principle is that each node should be given an equal slice of
> # the token ring; see http://wiki.apache.org/cassandra/Operations
> # for more details.
> #
> # If blank, Cassandra will request a token bisecting the range of
> # the heaviest-loaded existing node. If there is no load information
> # available, such as is the case with a new cluster, it will pick
> # a random token, which will lead to hot spots.
>
> Naturally, he is referring to that link when he says manual ;) AND notice the "which will lead to hot spots" comment.
>
> Later,
> Dean
>
> From: Darvin Denmian <da...@gmail.com>>
> Reply-To: "user@cassandra.apache.org<ma...@cassandra.apache.org>" <us...@cassandra.apache.org>>
> Date: Monday, October 1, 2012 6:56 AM
> To: "user@cassandra.apache.org<ma...@cassandra.apache.org>" <us...@cassandra.apache.org>>
> Subject: Re: Rebalancing cluster
>
> Did you set up the tokens the way
> the manual says?
Re: Rebalancing cluster
Posted by "Hiller, Dean" <De...@nrel.gov>.
You should check the cassandra.yaml file. There is an initial_token in that file that you should have set. The comment above that property reads
# You should always specify InitialToken when setting up a production
# cluster for the first time, and often when adding capacity later.
# The principle is that each node should be given an equal slice of
# the token ring; see http://wiki.apache.org/cassandra/Operations
# for more details.
#
# If blank, Cassandra will request a token bisecting the range of
# the heaviest-loaded existing node. If there is no load information
# available, such as is the case with a new cluster, it will pick
# a random token, which will lead to hot spots.
Naturally, he is referring to that link when he says manual ;) AND notice the "which will lead to hot spots" comment.
Later,
Dean
From: Darvin Denmian <da...@gmail.com>>
Reply-To: "user@cassandra.apache.org<ma...@cassandra.apache.org>" <us...@cassandra.apache.org>>
Date: Monday, October 1, 2012 6:56 AM
To: "user@cassandra.apache.org<ma...@cassandra.apache.org>" <us...@cassandra.apache.org>>
Subject: Re: Rebalancing cluster
Did you set up the tokens the way
the manual says?