Look in the logs for errors or warnings. Also let us know what version you are using.  

Am guessing that node 2 still thought that node 1 was in the cluster when you did the move. Which should(?) have errored. 


Aaron Morton
Freelance Developer

On 1/06/2012, at 1:50 AM, Poziombka, Wade L wrote:

Let me elaborate a bit.

two node cluster
node1 has token 0
node2 has token 85070591730234615865843651857942052864

node1 goes down perminently.

do a nodetool move 0 on node2.

monitor with ring... is in Moving state forever it seems.

From: Poziombka, Wade L
Sent: Tuesday, May 29, 2012 4:29 PM
To: user@cassandra.apache.org
Subject: nodetool move 0 gets stuck in "moving" state forever

If the node with token 0 dies and we just want it gone from the cluster we would do a nodetool move 0.  Then we monitor using nodetool ring it seems to be stuck on Moving forever.

Any ideas?