cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sanjeev Kulkarni <>
Subject Running multiple sstable loaders
Date Wed, 28 Mar 2012 21:57:23 GMT
Hey guys,
We have a fresh 4 node 0.8.10 cluster that we want to pump lots of data into.
The data resides on 5 data machines that are different from Cassandra
nodes. Each of these data nodes has 7 disks where the data resides.
In order to get maximum load performance, we are assigning 7 ips to
each data node on the same interface(eth2:0, eth2:1, ...). I also make
multiple copies of cassandra conf directory(/etc/cassandra). Each
cassandra conf is identical except for listen and rpc address. We then
start multiple simultaneous sstable loaders each pointing to different
config. We do this on all data nodes.
What we are seeing is the first sstableloader on each machine starts
loading. However the rest fail with connection refused error. They log
this message 8 times and then bail out. Any idea what could be wrong?

Sent from my iPhone

View raw message