You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@cassandra.apache.org by Utku Can Topçu <ut...@topcu.gen.tr> on 2010/10/04 11:58:51 UTC

Hardware change of a node in the cluster

Hey All,

Recently I've tried to upgrade (hw upgrade) one of the nodes in my cassandra
cluster from ec2-small to ec2-large.

However, there were problems and since the IP of the new instance was
different from the previous instance. The other nodes didnot recognize it in
the ring.

So what should be the best practice for a complete hardware change of one
node in the cluster while keeping the data that it has.

Regards,

Utku

Re: Hardware change of a node in the cluster

Posted by Gary Dusbabek <gd...@gmail.com>.
It should work this way:

1. Move your data to the new node (scp, etc.)
2. Make sure the new node is configured to use the same token as the new node.
3. Stand up the new node.
4. Turn off the old node.

If your environment is volatile, it's probably best to run `nodetool
repair` on the new node.

Gary.


On Mon, Oct 4, 2010 at 04:58, Utku Can Topçu <ut...@topcu.gen.tr> wrote:
> Hey All,
>
> Recently I've tried to upgrade (hw upgrade) one of the nodes in my cassandra
> cluster from ec2-small to ec2-large.
>
> However, there were problems and since the IP of the new instance was
> different from the previous instance. The other nodes didnot recognize it in
> the ring.
>
> So what should be the best practice for a complete hardware change of one
> node in the cluster while keeping the data that it has.
>
> Regards,
>
> Utku
>

Re: Hardware change of a node in the cluster

Posted by Jedd Rashbrooke <je...@imagini.net>.
On 4 October 2010 10:58, Utku Can Topçu <ut...@topcu.gen.tr> wrote:
> Recently I've tried to upgrade (hw upgrade) one of the nodes in my cassandra
> cluster from ec2-small to ec2-large.

 Something that bit me on this (I've done it with both
 Cassandra and Hadoop boxes, and some problems
 might be more Hadoopy related) is hostname.

 You need to change /etc/hostname and probably (unless
 you're happy to reboot again) the commandline of this
 too, to change it in the current instance.

 I found it best to drop all Cass instances (drain then stop),
 just to be a bit more confident.

 j.