Interestingly, if I run with no initialtoken specified, I get a "No
other nodes seen!" exception from the BootStrapper.  Full debug log
(minus rowmutations) here:
https://gist.github.com/df122c109bb9332cd85c

Thanks,

Andy

On Fri, Jul 2, 2010 at 1:23 AM, Andy Skalet <mailingli...@bitjug.com> wrote:
> Hello,
>
> I'm a new question asker here but I'll try to give you all the
> information you need.
> I'm trying to bootstrap a 3rd node into a two node cluster.  I am
> specifying an InitialToken of 0 for the new node.  I have replication
> factor set to 2, and am using the RackAwareStrategy with digg's
> PropertyFileEndPointSnitch.  I'm using an /etc/hosts file to map ips
> to short hostnames which the interface bindings and seeds in the
> config files use.   Here is what the ring looks like before I attempt
> to bootstrap the new node:
>
> $ apache-cassandra-0.6.3/bin/nodetool -h localhost ring
> Address       Status     Load          Range
>           Ring
>
> 138705532905298861485169149262903205000
> 10.198.7.143  Up         4.57 GB
> 85070591730234615865843651857942052864     |<--|
> 10.211.71.241 Up         7.13 GB
> 138705532905298861485169149262903205000    |-->|
>
> This is with cassandra 0.6.3 obviously.  The exception I get is:
>
> java.lang.AssertionError
>        at 
> org.apache.cassandra.dht.BootStrapper.getRangesWithSources(BootStrapper.java:146)
>        at 
> org.apache.cassandra.dht.BootStrapper.startBootstrap(BootStrapper.java:72)
>        at 
> org.apache.cassandra.service.StorageService.startBootstrap(StorageService.java:406)
>        at 
> org.apache.cassandra.service.StorageService.initServer(StorageService.java:358)
>        at 
> org.apache.cassandra.thrift.CassandraDaemon.setup(CassandraDaemon.java:115)
>        at 
> org.apache.cassandra.thrift.CassandraDaemon.main(CassandraDaemon.java:211)
>
> I've posted the full debug log, minus the verbose row mutation messages, here:
>
> https://gist.github.com/11fc9eddd524e269e490
>
> I see in the code that the tokenMetadata_ is supposed to have a
> non-zero number of tokens in it, which it looks like should be created
> by updateNormalToken.  I have no idea why this isn't happening in my
> case, however.
>
> Thanks in advance for any ideas,
>
> Andy Skalet
>

Reply via email to