You might need to do the nodetool scrub on the nodes to rebuild the sstables for the different protocols.
On Fri, Sep 16, 2011 at 10:50 PM, Yan Chunlu <springri...@gmail.com> wrote: > and also the load is unusual(node1 has 80M data before the upgrade): > > bash-3.2$ bin/nodetool -h localhost ring > Address DC Rack Status State Load Owns > Token > > 93798607613553124915572813490354413064 > node2 datacenter1 rack1 Up Normal 86.03 MB 46.81% > 3303745385038694806791595159000401786 > node3 datacenter1 rack1 Up Normal 67.68 MB 26.65% > 48642301133762927375044585593194981764 > node1 datacenter1 rack1 Up Normal 114.81 KB 26.54% > 93798607613553124915572813490354413064 > > > > On Fri, Sep 16, 2011 at 10:48 PM, Yan Chunlu <springri...@gmail.com>wrote: > >> after kill node1 and start it again, node 3 has the same problems with >> node2... >> >> >> On Fri, Sep 16, 2011 at 10:42 PM, Yan Chunlu <springri...@gmail.com>wrote: >> >>> I am running local tests about upgrade cassandra. upgrade from 0.7.4 to >>> 0.8.5 >>> after upgrade one node1, two problem happened: >>> >>> 1, node2 keep saying: >>> >>> "Received connection from newer protocol version. Ignorning message." >>> >>> is that normal behaviour? >>> >>> 2, while running "describe cluster" on node1, it shows node2 unreachable: >>> Cluster Information: >>> Snitch: org.apache.cassandra.locator.SimpleSnitch >>> Partitioner: org.apache.cassandra.dht.RandomPartitioner >>> Schema versions: >>> UNREACHABLE: [node2] >>> 05f1ee3b-e063-11e0-97d5-63c2fb3f0ca8: [node1, node3] >>> >>> node3 seems act normal. >>> >>> >>> I saw the JMXPORT has changed since 0.8, is that the reason node was >>> unreachable? >>> >>> >>> thanks! >>> >>> >>> >> > -- Dikang Gu 0086 - 18611140205
