CODEWITHSUNDEEP
cassandracql
Pradyumn
Pradyumn

Reputation: 430

Cassandra, strange behavior after changing replication_factor

I am facing a strange situation when changing the replication_factor of an existing keyspace.

Using a cluster of three nodes, each running Datastax Cassandra community edition 2.0.11.

I created a keyspace 'training':

create keyspace training with replication = {
'class': 'SimpleStrategy',
'replication_factor': 1
};

Then,

use training;

Then, 

create table customers (...);

Inserted three records. Queried the table from another node, and got the three records back, as expected.

Then, changed the replication_factor:

alter keyspace training with replication = {
'class': 'SimpleStrategy',
'replication_factor': 3
};

Queried the table. It showed zero records! After a few seconds, it showed 2 records. Then again zero records. Inserted two records. Now it sometimes shows 2 (new) records, sometimes 4 (2 new, 2 old). Same behavior on all three machines.

I am absolutely stumped. Does anybody have a clue why this is behaving this way?

Only one weird possibility comes to my mind: out of three nodes in the cluster, two are running Ubuntu, 64-bit OS, and the third is running 32-bit Windows 7 OS. Does that matter / explain this behavior?

Thanks.

Upvotes: 3

Views: 194

Answers (1)

Pradyumn
Pradyumn

Reputation: 430

Problem solved. I realized that I had not run "nodetool repair". Once I ran that the mysterious disappearance / reapparance stopped.

Upvotes: 4

Related Questions