After restarting all the nodes, they all see each other. I'll try nodetool gossipinfo if it happens again. Thanks.
From: Aaron Morton [mailto:aa...@thelastpickle.com] Sent: Monday, December 23, 2013 10:19 PM To: Cassandra User Subject: Re: Issue upgrading from 1.2 to 2.0.3 If this is still a concern can you post the output from nodetool gossipinfo ? It will give the details of the nodes think of the other ones. A ----------------- Aaron Morton New Zealand @aaronmorton Co-Founder & Principal Consultant Apache Cassandra Consulting http://www.thelastpickle.com On 20/12/2013, at 11:38 am, Parag Patel <parag.pa...@fusionts.com<mailto:parag.pa...@fusionts.com>> wrote: Thanks for that link. Our 1.2 version is 1.2.12 Our 2.0.3 nodes were restarted once. Before restart, it was the 1.2.12 binary, after it was the 2.0.3. Immediately after the node was back in the cluster, we ran nodetool upgradesstables. We haven't restarted since. Is a restart required for each node? From: Robert Coli [mailto:rc...@eventbrite.com] Sent: Thursday, December 19, 2013 4:17 PM To: user@cassandra.apache.org<mailto:user@cassandra.apache.org> Subject: Re: Issue upgrading from 1.2 to 2.0.3 On Thu, Dec 19, 2013 at 1:03 PM, Parag Patel <parag.pa...@fusionts.com<mailto:parag.pa...@fusionts.com>> wrote: We are in the process of upgrading 1.2 to 2.0.3. ... Please help as this will prevent us from pushing into production. (as a general commentary : https://engineering.eventbrite.com/what-version-of-cassandra-should-i-run/ ) specific feedback on your question : Did the 2.0.3 nodes see the 1.2.x (which 1.2.x?) nodes after the first restart? =Rob