Replace dead node in non-vnode 1.2 cluster

[John Wong <go...@gmail.com>]

Hi

Can the community help to confirm that
http://docs.datastax.com/en/cassandra/1.2/cassandra/operations/ops_replace_node_t.html
will work for non-vnode cluster in Cassandra 1.2.

It looks like I don't have to set the initial token for the replacement
node (but same IP) at all if I run the JVM opts during startup.

I read a couple mailing list threads.

[1]:
http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Bootstrap-failure-on-C-1-2-13-td7592664.html

[2]:
http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Replacing-a-dead-node-in-Cassandra-2-0-8-td7596245.html

They mentioned that back in the date 1.2 doc would defer user to 1.1 doc,
but the current 1.2 doc doesn't mention anything about that.

Also these as references

[3]:
https://wiki.apache.org/cassandra/Operations#For_versions_1.2.0_and_above
[4]:
http://blog.alteroot.org/articles/2014-03-12/replace-a-dead-node-in-cassandra.html

Thank you.

John

Re: Replace dead node in non-vnode 1.2 cluster

[John Wong <go...@gmail.com>]

Hi Paulo.
Thank you. Since you created the original thread in one of my references,
we are going to do with that instructions. Should be okay :-)

John

On Thu, Sep 17, 2015 at 10:48 PM, Paulo Motta <pa...@gmail.com>
wrote:

> The datastax documentation was fixed after the initial confusion with
> vnodes vs non-vnodes, so you should be safe to follow the procedure
> described there. Make sure to set the
> -Dcassandra.replace_address=address_of_dead_node JVM option (don't worry
> about the initial token).
>
> 2015-09-17 21:21 GMT-03:00 John Wong <go...@gmail.com>:
>
>> Hi
>>
>> Can the community help to confirm that
>> http://docs.datastax.com/en/cassandra/1.2/cassandra/operations/ops_replace_node_t.html
>> will work for non-vnode cluster in Cassandra 1.2.
>>
>> It looks like I don't have to set the initial token for the replacement
>> node (but same IP) at all if I run the JVM opts during startup.
>>
>> I read a couple mailing list threads.
>>
>> [1]:
>> http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Bootstrap-failure-on-C-1-2-13-td7592664.html
>>
>> [2]:
>> http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Replacing-a-dead-node-in-Cassandra-2-0-8-td7596245.html
>>
>> They mentioned that back in the date 1.2 doc would defer user to 1.1 doc,
>> but the current 1.2 doc doesn't mention anything about that.
>>
>> Also these as references
>>
>> [3]:
>> https://wiki.apache.org/cassandra/Operations#For_versions_1.2.0_and_above
>> [4]:
>> http://blog.alteroot.org/articles/2014-03-12/replace-a-dead-node-in-cassandra.html
>>
>> Thank you.
>>
>> John
>>
>
>

Re: Replace dead node in non-vnode 1.2 cluster

[Paulo Motta <pa...@gmail.com>]

The datastax documentation was fixed after the initial confusion with
vnodes vs non-vnodes, so you should be safe to follow the procedure
described there. Make sure to set the
-Dcassandra.replace_address=address_of_dead_node JVM option (don't worry
about the initial token).

2015-09-17 21:21 GMT-03:00 John Wong <go...@gmail.com>:

> Hi
>
> Can the community help to confirm that
> http://docs.datastax.com/en/cassandra/1.2/cassandra/operations/ops_replace_node_t.html
> will work for non-vnode cluster in Cassandra 1.2.
>
> It looks like I don't have to set the initial token for the replacement
> node (but same IP) at all if I run the JVM opts during startup.
>
> I read a couple mailing list threads.
>
> [1]:
> http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Bootstrap-failure-on-C-1-2-13-td7592664.html
>
> [2]:
> http://cassandra-user-incubator-apache-org.3065146.n2.nabble.com/Replacing-a-dead-node-in-Cassandra-2-0-8-td7596245.html
>
> They mentioned that back in the date 1.2 doc would defer user to 1.1 doc,
> but the current 1.2 doc doesn't mention anything about that.
>
> Also these as references
>
> [3]:
> https://wiki.apache.org/cassandra/Operations#For_versions_1.2.0_and_above
> [4]:
> http://blog.alteroot.org/articles/2014-03-12/replace-a-dead-node-in-cassandra.html
>
> Thank you.
>
> John
>