Make sure you test it on a practice cluster. Messing with the system tables
is risky business!

raft.so - Cassandra consulting, support, and managed services


On Sun, Feb 21, 2021 at 11:12 PM Paul Chandler <p...@redshots.com> wrote:

> Hi Kane,
>
> That sounds a good idea, I will give it a try on Monday.
>
> Thanks
>
> Paul
>
> On 21 Feb 2021, at 11:33, Kane Wilson <k...@raft.so> wrote:
>
> There has been proposals to add a force/unsafe flag to alter DC but it
> hasn't been actioned and at this rate seems unlikely to make it into 4.0.
> There is however a workaround, albeit not very user friendly. You should be
> able to modify the system_schema tables directly to do your DC updates. I
> am yet to test it out myself so don't know the updates you'd need to make
> but you should be able to get a good idea by querying, doing updates and
> observing the effect.
>
> raft.so - Cassandra consulting, support, managed services
>
> On Sat., 20 Feb. 2021, 02:29 Paul Chandler, <p...@redshots.com> wrote:
>
>> All,
>>
>> We have a use case where we need to change the datacenter name for a
>> cassandra cluster, we have a script to do this that involves a short
>> downtime. This does the following
>>
>>
>> 1) Change replication factor for the system key spaces to be { ‘OLD_DC’ :
>> ‘3’, ’NEW_DC”: ‘3’  }
>> 2) Change the dc value in cassandra-rackdc.properties to NEW_DC for each
>> node
>> 3) Add -Dcassandra.ignore_dc=true in cassandra-env.sh for each node
>> 4) Stop all nodes
>> 5) Start each seed node, then start the rest of the nodes
>> 6) change the replication factor for all the keyspaces to {  ’NEW_DC”:
>> ‘3’  }
>>
>> In 3.11.x this all works fine and the cluster is good to be used again
>> after step 6.
>>
>> However in 4.0 step 1 is now blocked by the following change "Cassandra
>> will no longer allow invalid keyspace replication options, such as invalid
>> datacenter names for NetworkTopologyStrategy”
>>
>> If you skip step 1) then when the nodes come back up, you cannot login
>> because the system_auth keyspace still has a replication factor of ‘OLD_DC’
>> : ‘3’ but there are no nodes in the dc OLD_DC so the keyspace cannot be
>> accessed.
>>
>> Is there a way around this to change the name of the datacenter?
>>
>>
>> Thanks
>>
>> Paul
>>
>>
>>
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: user-unsubscr...@cassandra.apache.org
>> For additional commands, e-mail: user-h...@cassandra.apache.org
>>
>>
>

Reply via email to