Yes, at the moment it is not possible to switch from Random to Murmur.
We were interested to move to Murmur. And we looked if it was possible when
migrating the keyspaces to another datacenter. However this is not possible
even while the new datacenter didn’t have the same number of nodes.
-- Bri
(forgot to mention, if there's more to add, please move this over to the
standard cassandra mailing list -- this list is used for development of
cassandra itself).
RandomPartitioner will probably be there for a long time for backwards
compat, and if it goes away (unlikely) there would certainly be
We have running clusters which use RandomPartitioner that's why I wonder if we
could use this feature in the future.
Best,
Romain
RandomPartitioner is included for backwards compatibility, though not
suggested for new projects. As far as I know Murmur is faster and offers
less hotspots in token allocation.
Cheers,
Russ
On Wed, Feb 10, 2016 at 10:04 AM, Romain Hardouin
wrote:
> Hi all,
>
> cassandra.yaml mentions that the
Hi all,
cassandra.yaml mentions that the new token allocation algorithm is only
supported with the Murmur3Partitioner [1].
When looking at the commit [2] I see that IPartitioner interface is used.
I do see references to Murmur3Partitioner but only in tests classes.
What would prevent to use Rand
Take a look at MetadataSerializer and the MetadataComponent subclasses.
On Tue, Feb 9, 2016 at 4:34 PM, Rajath Subramanyam
wrote:
> Hello Cassandra-Dev,
>
> I have noticed that in Cassandra 3.0 there is a new file in the
> // called
> ma--big-Statistics.db.
>
> What does this file contain ? Is i
