What causes "Failed to retrieve compression dictionary for {}.{}. {}" in apache/cassandra?

Verify the dictionary row exists in system_distributed (dictionary table) for the given table id and dictionary id.

Failed to retrieve compression dictionary for

Error message

Failed to retrieve compression dictionary for {}.{}. {}

What it means

The CompressionDictionaryEventHandler, on learning a new dictionary is available, tries to read it from the system_distributed keyspace and add it to the local cache. If retrieval fails (missing row, deserialization issue, I/O error) it logs this warning and continues without the dictionary; compression falls back to operating without that dictionary.

Solutions

Verify the dictionary row exists in system_distributed (dictionary table) for the given table id and dictionary id.
Run repair on system_distributed so the dictionary data replicates to this node, then wait for the scheduler to retry.
Check connectivity/consistency to system_distributed and the logged exception for the root cause.
Disable/re-enable dictionary compression for the table to force a fresh training/retrieval cycle.

Defensive patterns

Strategy: retry

Validate before calling

// verify the dictionary row exists before assuming failure is fatal
Row r = session.execute("SELECT * FROM system_distributed.compression_dictionaries WHERE keyspace_name=? AND table_name=?",
                        ks, table).one();

Try / catch

// the handler already swallows; operators retry via repair + next scheduled refresh
try { retrieveDictionary(); } catch (Exception e) { scheduleRetryWithBackoff(e); }

Prevention

Keep system_distributed repaired, especially after adding nodes.
Do not delete dictionary rows before peers have pulled them.

When it happens

Trigger: onNewDictionaryAvailable runs and SystemDistributedKeyspace.retrieveCompressionDictionary throws or returns null for the given keyspace/table/dictionaryId — e.g. the dictionary row was not yet replicated locally.

Common situations: New node pulling dictionary metadata before full repair/replication of system_distributed; network or consistency issues reading system_distributed; dictionary deleted on the writer side before peers fetched it.

Understand the failure class

Background: 'Could not be found', 'does not exist', 'not found in database': the resource-not-found family when an ID, slug, key, or URI lookup comes back empty — this error's family across 20 libraries.

Related errors

AI-assisted analysis of apache/cassandra@88fd0f6a0e (2026-09-10). Data as JSON: /api/errors/4f681b17d868d0e4. Report an issue: GitHub.

Appendix: source

Thrown at src/java/org/apache/cassandra/db/compression/CompressionDictionaryEventHandler.java:95

    @Override
    public void onNewDictionaryAvailable(CompressionDictionary.DictId dictionaryId)
    {
        // Best effort to retrieve the dictionary; otherwise, the periodic task should retrieve the dictionary later
        ScheduledExecutors.nonPeriodicTasks.submit(() -> {
            try
            {
                if (!cfs.metadata().params.compression.isDictionaryCompressionEnabled())
                {
                    return;
                }

                CompressionDictionary dictionary = SystemDistributedKeyspace.retrieveCompressionDictionary(keyspaceName, tableName, cfs.metadata().id.toLongString(), dictionaryId.id);
                cache.add(dictionary);
            }
            catch (Exception e)
            {
                logger.warn("Failed to retrieve compression dictionary for {}.{}. {}",
                            keyspaceName, tableName, dictionaryId, e);
            }
        });
    }

    // Best effort to notify the peer regarding the new dictionary being available to pull.
    // If the request fails, each peer has periodic task scheduled to pull.
    private void sendNotification(InetAddressAndPort target, CompressionDictionaryUpdateMessage message)
    {
        logger.debug("Sending dictionary update notification for {} to {}", message.dictionaryId, target);

        Message<CompressionDictionaryUpdateMessage> msg = Message.out(Verb.DICTIONARY_UPDATE_REQ, message);
        MessagingService.instance()
                        .sendWithResponse(target, msg)
                        .addListener(future -> {
                            if (future.isSuccess())
                            {
                                logger.debug("Successfully sent dictionary update notification to {}", target);

View on GitHub (pinned to 88fd0f6a0e)