Unable to do rolling upgarde from 1.8.1

Hi,

I have been trying to upgrade a 1.8.1 cluster to 2.0.0 but have been unsuccessful. Every time the rebalance operation after adding the first 2.0.0 node fails. Here are the steps I have followed:

1. Started with a 3 node cluster of 1.8.1 (CentOS 6.2, 64-bit).
2. Removed node 1 from the cluster
3. Rebalanced the cluster (successful)
4. Uninstalled couchbase 1.8.1 from the node 1
5. deleted /opt/couchbase folder
6. Installed couchbase 2.0.0
7. Logged into web console of node 1
8. Joined existing cluster and clicked on rebalance.
9. Nothing seems to happen for a good 5 odd minutes. And then the logs show the following error:

2012-12-20 04:46:13.417 ns_orchestrator:2:info:message(ns_1@172.28.136.35) - Rebalance exited with reason {not_all_nodes_are_ready_yet,
['ns_1@172.28.136.35']}

I have tried this 3 times. Every time the same result. BTW, .35 is node1 (as referred above)

Is there a wait time or something after installing couchbase 2.0.0 on a node?

Any other insights?

regards,
-Piyush