For ease of use, we actually had a single cassandra.yaml deployed to every machine and a script that swapped out the token and listen address. I had seed nodes ip1,ip2,ip3 as the seeds but what I didn't realize was then that these nodes had themselves as seeds. I am assuming that should never be done, is that correct. I really should deploy ip1, ip2, ip3 on all nodes and then for nodes 1, 2, and 3, I should do something like
ip1 will have ip2, ip3, ip4 ip2 will have ip1, ip3, ip4 Etc. etc. QUESTION: Would it be ok if I just configured every node to be the 3 ip's after it like this instead as well ip1 would have ip2, ip3, ip4 ip2 would have ip3, ip4, ip5 ip3 would have ip4, ip5, ip6 ip4 would have ip5, ip6, ip1 Is this okay for seed node configuration? Thanks, Dean
