apache/cassandra · error · ConfigurationException

Allocated token %s already assigned to node %s. Is another n

Error message

Allocated token %s already assigned to node %s. Is another node also allocating tokens?

What it means

TokenAllocation throws ConfigurationException when a token chosen by the allocator is already owned by another node in the same allocation ring (same datacenter/rack scope). It indicates two nodes are allocating tokens concurrently or a racing allocation has claimed the same token.

Source

Thrown at src/java/org/apache/cassandra/dht/tokenallocator/TokenAllocation.java:187

                InetAddressAndPort endpoint = metadata.directory.endpoint(en.getValue());
                if (inAllocationRing(metadata.locator, endpoint))
                    sortedTokens.put(en.getKey(), endpoint);
            }
            return TokenAllocatorFactory.createTokenAllocator(sortedTokens, this, metadata.tokenMap.partitioner());
        }

        final Collection<Token> adjustForCrossDatacenterClashes(Collection<Token> tokens)
        {
            List<Token> filtered = Lists.newArrayListWithCapacity(tokens.size());

            for (Token t : tokens)
            {
                while (metadata.tokenMap.owner(t) != null)
                {
                    NodeId nodeId = metadata.tokenMap.owner(t);
                    InetAddressAndPort other = metadata.directory.endpoint(nodeId);
                    if (inAllocationRing(metadata.locator, other))
                        throw new ConfigurationException(String.format("Allocated token %s already assigned to node %s. Is another node also allocating tokens?", t, other));
                    t = t.nextValidToken();
                }
                filtered.add(t);
            }
            return filtered;
        }

        final SummaryStatistics replicatedOwnershipStats()
        {
            SummaryStatistics stat = new SummaryStatistics();
            for (Map.Entry<InetAddressAndPort, Double> en : evaluateReplicatedOwnership().entrySet())
            {
                // Filter only in the same allocation ring
                if (inAllocationRing(metadata.locator, en.getKey()))
                {
                    NodeId nodeId = metadata.directory.peerId(en.getKey());
                    stat.addValue(en.getValue() / metadata.tokenMap.tokens(nodeId).size());
                }

View on GitHub (pinned to 88fd0f6a0e)

Solutions

1. Retry the node bootstrap after the other allocation completes (serialize node joins)
2. Ensure only one node allocates tokens at a time — use deterministic ordering in provisioning scripts
3. Check nodetool ring for the conflicting token owner and choose a different token/bootstrap order
4. If metadata is stale, wait for gossip convergence and retry

Example fix

// before: two nodes bootstrap in parallel
ansible all -m cassandra_bootstrap  # race -> ConfigurationException
// after: join nodes sequentially
- hosts: cassandra[0]
  tasks: [bootstrap node]
- hosts: cassandra[1]
  tasks: [bootstrap next node after previous is Up/Normal]

// before bootstrapping a new node, confirm no other node is currently bootstrapping
nodetool status  # no 'UJ' (Up/Joining) nodes
// and confirm the candidate tokens are unowned in nodetool ring

try {
    TokenAllocation.allocate(metadata, ks, dc);
} catch (ConfigurationException e) {
    if (e.getMessage().contains("already assigned")) {
        // backoff and retry once other node finished joining
    }
}

When it happens

Trigger: Calling allocate when, during adjustForCrossDatacenterClashes, metadata.tokenMap.owner(token) returns a non-null node that is inAllocationRing — i.e. the token is taken by another node in the same DC.

Common situations: Two or more nodes bootstrapping simultaneously with token allocation enabled; a node bootstrapping while another allocation just finished; clustered automated provisioning racing across nodes.

Related errors

• Queue is empty
• UnsupportedOperationException
• Maximum pool size has been changed while resizing
• Must be one of ${values}
• ${executor} has shut down

AI-assisted analysis of apache/cassandra@88fd0f6a0e (2026-09-10). Data as JSON: /api/errors/dad6ec3433ec2ccc. Report an issue: GitHub.