[CASSANDRA-1924] Broken keyspace strategy_option with zero replicas

• Type: Bug
• Status: Resolved
• Priority: Low
• Resolution: Fixed
• Fix Version/s: 0.7.4
• Severity: Low
• Assignee: Sylvain Lebresne
• Reporter: Thor Carpenter

Authors

Sylvain Lebresne

Reviewers

Jonathan Ellis

• Created: 30/Dec/10 22:59
• Updated: 16/Apr/19 09:33
• Resolved: 11/Mar/11 18:18

Description

When a keyspace is defined that has strategy options specifying zero replicas should be place in a datacenter (e.g. DCX:0), an assert is violated for any insert and LOCAL_QUORUM reads fail. I'm not sure if the issue is that there are no nodes in DCX or that I'm saying DCX shouldn't get any replicas, or a combination of the two.

The broken keyspace:

create keyspace KeyspaceDC1 with
replication_factor = 1 and
placement_strategy = 'org.apache.cassandra.locator.NetworkTopologyStrategy' and
strategy_options = [

{DC1:1, DC2:0}

];

The fixed keyspace:

create keyspace KeyspaceDC1 with
replication_factor = 1 and
placement_strategy = 'org.apache.cassandra.locator.NetworkTopologyStrategy' and
strategy_options = [

{DC1:1}

];

To reproduce:

• Install the 0.7rc3 rpm on a single node in "DC1".
• In cassandra.yaml set initial_token = 1 and specify PropertyFileSnitch.
• cassandra-topology.properties:

10.5.64.26=DC1:R1
default=DC2:R1

• Schema loaded via cassandra-cli:

create keyspace KeyspaceDC1 with
replication_factor = 1 and
placement_strategy = 'org.apache.cassandra.locator.NetworkTopologyStrategy' and
strategy_options = [

{DC1:1, DC2:0}

];

use KeyspaceDC1;

create column family TestCF with
column_type = 'Standard' and
comparator = 'BytesType' and
keys_cached = 200000 and
rows_cached = 2000 and
gc_grace = 0 and
read_repair_chance = 0.0;

• In cassandra-cli execute the following:

[default@unknown] use KeyspaceDC1;
Authenticated to keyspace: KeyspaceDC1
[default@KeyspaceDC1] set TestCF['some key']['some col'] = 'some value';
Internal error processing insert

• If you have asserts enabled, check system.log where you should find the assertion error:

DEBUG [pool-1-thread-3] 2010-12-29 12:10:38,897 CassandraServer.java (line 362) insert
ERROR [pool-1-thread-3] 2010-12-29 12:10:38,906 Cassandra.java (line 2960) Internal error processing insert
java.lang.AssertionError
at org.apache.cassandra.locator.TokenMetadata.firstTokenIndex(TokenMetadata.java:392)
at org.apache.cassandra.locator.TokenMetadata.ringIterator(TokenMetadata.java:417)
at org.apache.cassandra.locator.NetworkTopologyStrategy.calculateNaturalEndpoints(NetworkTopologyStrategy.java:95)
at org.apache.cassandra.locator.AbstractReplicationStrategy.getNaturalEndpoints(AbstractReplicationStrategy.java:99)
at org.apache.cassandra.service.StorageService.getNaturalEndpoints(StorageService.java:1411)
at org.apache.cassandra.service.StorageService.getNaturalEndpoints(StorageService.java:1394)
at org.apache.cassandra.service.StorageProxy.mutate(StorageProxy.java:109)
at org.apache.cassandra.thrift.CassandraServer.doInsert(CassandraServer.java:442)
at org.apache.cassandra.thrift.CassandraServer.insert(CassandraServer.java:379)
at org.apache.cassandra.thrift.Cassandra$Processor$insert.process(Cassandra.java:2952)
at org.apache.cassandra.thrift.Cassandra$Processor.process(Cassandra.java:2555)
at org.apache.cassandra.thrift.CustomTThreadPoolServer$WorkerProcess.run(CustomTThreadPoolServer.java:167)
at java.util.concurrent.ThreadPoolExecutor$Worker.runTask(ThreadPoolExecutor.java:886)
at java.util.concurrent.ThreadPoolExecutor$Worker.run(ThreadPoolExecutor.java:908)
at java.lang.Thread.run(Thread.java:619)

• If you don't have asserts enabled, you should find that no errors are logged but LOCAL_QUORUM reads cause TimedOutExceptions on the client.