amazon ec2 - nodetool status in cassandra giving previous IP of nodes -
i trying expand 2 node cassandra cluster 4 node cluster 4 amazon ec2 instances. have created 4 nodes , made following changes in cassandra.yaml file.
listen_address = 10.30.143.145 seeds = 10.30.143.145,10.159.58.234,10.170.31.252,10.158.52.84 endpoint_snitch: ec2snitch num_tokens: 256
i have replicated these changes across 4 nodes. had expanded single node cluster double node cluster following procedure. however, after configuring 4 node cluster, when ./nodetool status on first node, following output:
ubuntu@ip-10-170-31-252:~/viq-cloud/software/apache-cassandra-1.2.5/bin$ ./nodetool status datacenter: us-east =================== status=up/down |/ state=normal/leaving/joining/moving -- address load tokens owns host id rack un 10.30.143.145 927.14 kb 256 16.1% 34d0424a-fe07-4047-a2a5-f45b9a0049d6 1a un 10.159.58.234 135.2 kb 256 15.7% 00308009-8755-4bce-906f-4eda53a31fc6 1a un 10.170.31.252 20.94 gb 256 20.9% a815f0de-64db-418c-97a3-9aa7be280280 1a un 10.158.52.84 311.33 kb 256 15.0% fc634f65-3cf3-4e24-a9a3-456adbd174e0 1a datacenter: unknown-dc ====================== status=up/down |/ state=normal/leaving/joining/moving -- address load tokens owns host id rack dn 10.170.22.214 ? 256 16.6% 253ee376-c49d-47a1-a321-4f155870c122 unknown-rack dn 10.31.131.35 ? 256 15.7% e94c0cb1-9635-42c9-8982-450271f7da1c unknown-rack
the address of 2 nodes shown dn of previous private ip's of other ec2 nodes (each reboot of ec2 instance changes private ips).
other 3 nodes giving me proper result nodetool status. wondering cassandra picking previous ip's because , havent mentioned them in cassandra.yaml file.
i followed instructions here adding new nodes. please advise on why happening.
if haven't figured out, think answer may need make sure using ec2 or ec2multiregion snitch in cassandra.yaml on new nodes. fact got put in "unknown-dc" means existing nodes don't know dc put them in.
Comments
Post a Comment