0

I am using cassandra dse 4.7, 3 node running cluster with one data center, now we are planning to add another data center with 3 nodes. I followed the steps mentioned in:
http://docs.datastax.com/en/cassandra/2.0/cassandra/operations/ops_add_dc_to_cluster_t.html.
We have around 20GB of data on each node (total : 60GB).

The problem is when I run: 

nodetool rebuild --old_datacenter_name on one of new node

it is very slow to sync data from cluster. I have already increased getstreamthroughput and getcompactionthroughput but still no luck.

Are there any other parameters I could tune for faster synchronisation of the data to a newly added node?

HBruijn
  • 72,524
  • 21
  • 127
  • 192
Rajnish
  • 101
  • 3

1 Answers1

0

I found the reason why it is slow and crashed my node during rebuild ,
in my case the problem is secondary index , I created secondary index on each table and during rebuild it is more time consuming and eating all heap memory , due to this my node crashed many times .

I just delete secondary index and tried it is working fine. after rebuild I created secondary index again .

Rajnish
  • 101
  • 3