Hi All
We are running Cassandra 0.6.3,We have two node's with replication
factor one and ordered partitioning.Problem we are facing at the moment
all data is being send to one Cassandra node and its filling up quite
rapidly and we are short of disk space.Unfortunately we have hardware
With the Order Preserving Partitioner you are responsible for balancing the
rows around the cluster,
http://wiki.apache.org/cassandra/Operations?highlight=%28partitioner%29#Token_selection
Was there a reason for using the ordered partitioner rather than the random
one?
What does the
Below is the ouput of nodetool ring
Address Status Load
Range Ring
TuL8jLqs7uxLipP6
192.168.100.3 Up 89.91 GB
JDtVOU0YVQ6MtBYA |--|
192.168.100.4 Up
Dear Aaron,
We are little confused about OPP token.How to calculate OPP Token? Few
of our column families have UUID as key and other's have integer as key.
On 03/17/2011 04:22 PM, Ali Ahsan wrote:
Below is the ouput of nodetool ring
Address Status Load
Range
Please can any one give their comment on this
On 03/17/2011 07:02 PM, Ali Ahsan wrote:
Dear Aaron,
We are little confused about OPP token.How to calculate OPP Token? Few
of our column families have UUID as key and other's have integer as key.
From OrderPreservingPartition.java
public StringToken getToken(ByteBuffer key)
{
String skey;
try
{
skey = ByteBufferUtil.string(key, Charsets.UTF_8);
}
catch (CharacterCodingException e)
{
throw new