incubator-cassandra-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From aaron morton <aa...@thelastpickle.com>
Subject Re: Cassandra multi-datacenter
Date Mon, 06 May 2013 08:56:18 GMT
The broadcast_address can be set manually without using the EC2MultiRegionSnitch. It's the
address the node wants other nodes to talk to it on http://www.datastax.com/docs/1.2/configuration/node_configuration#broadcast-address
 
You may find it easier to run a VPN between the colo nodes and the EC2 nodes.

Cheers
 
-----------------
Aaron Morton
Freelance Cassandra Consultant
New Zealand

@aaronmorton
http://www.thelastpickle.com

On 4/05/2013, at 5:25 AM, Daning Wang <daning@netseer.com> wrote:

> Thanks Jabbar and Aaron.
> 
> Aaron - for broadcast_address , looks it is only working with EC2MultiRegionSnitch. but
in our case, we will have one center in colo, and one center in ec2(sorry, did not make that
clear, we'd like to replicate data from colo to EC2)
> 
> So can we still use broadcast_address? or other solutions? Is that easy to write a new
Snitch for this?
> 
> Thanks,
> 
> Daning
> 
> 
> 
> On Thu, May 2, 2013 at 2:31 PM, aaron morton <aaron@thelastpickle.com> wrote:
> Look at the broadcast_address in the yaml file
> Cheers
> -----------------
> Aaron Morton
> Freelance Cassandra Consultant
> New Zealand
> 
> @aaronmorton
> http://www.thelastpickle.com
> 
> On 3/05/2013, at 9:10 AM, Jabbar Azam <ajazam@gmail.com> wrote:
> 
>> I'm not sure why you want to use public Ip's in the other data centre. You're cassandra
nodes in the other datacentre will be accessible from the internet
>> 
>> Personally I would use private IP addresses in the second data centre, on a different
IP subnet.
>> 
>> A VPN is your only solution if you want to keep your data private and unhackable,
as it's tunneling it's way through the internet
>> 
>> A slow network connection will mean your data is not in sync in both datacentres
unless you explicitly specify quorum as your consisteny level in your mutation requests but
your database throughput will be affected by this.
>> 
>> You bandwidth to the second datacentre and the quantity of your mutation requests
will dictate how long it will take the second datacentre to get in sync with the primary datacentre.
>> 
>> 
>> I've probably missed something but there are plenty of intelligent people in this
mailing list to fill the blanks :)
>> 
>> Thanks
>> 
>> Jabbar Azam
>> 
>> 
>> On 2 May 2013 20:28, Daning Wang <daning@netseer.com> wrote:
>> Hi all,
>> 
>> We are deploying Cassandra on two data centers. there is slower network connection
between data centers. 
>> 
>> Looks casandra should use internal ip to communicate with nodes in the same data
center, and public ip to talk to nodes in other data center. We know VPN is a solution, but
want to know if there is other idea.
>> 
>> Thanks in advance,
>> 
>> Daning
>> 
> 
> 


Mime
View raw message