> What is the good practice to put in the code as addContactPoint ie.,how many > servers ? I use the same nodes as the seed list nodes for that DC.
The idea of the seed list is that it’s a list of well known nodes, and it’s easier operationally to say we have one list of well known nodes that is used by the servers and the clients. > 1) I am also thinking to put this way here I am not sure this good or bad > if i conigure 4 serves into one VIP ( virtual IP/virtual DNS) > and specifying that DSN in the code as ContactPoint, so that that VIP is > smart enough to route to different nodes. Too complicated. > 2) Is that problem if i use multiple Data centers in future ? You only need to give the client the local seeds, it will discover all the nodes. Cheers ----------------- Aaron Morton New Zealand @aaronmorton Co-Founder & Principal Consultant Apache Cassandra Consulting http://www.thelastpickle.com On 7/12/2013, at 7:12 am, chandra Varahala <hadoopandcassan...@gmail.com> wrote: > Greetings, > > I have 4 node cassandra cluster that will grow upt to 10 nodes,we are using > CQL Java client to access the data. > What is the good practice to put in the code as addContactPoint ie.,how many > servers ? > > 1) I am also thinking to put this way here I am not sure this good or bad > if i conigure 4 serves into one VIP ( virtual IP/virtual DNS) > and specifying that DSN in the code as ContactPoint, so that that VIP is > smart enough to route to different nodes. > > 2) Is that problem if i use multiple Data centers in future ? > > > thanks > Chandra