Support for Storage Attached Indexes (SAI)

Storage-Attached Indexing (SAI) completed its Beta and is now a generally available (GA) DSE feature. Use it to create one or more secondary indexes on DSE tables. See (SAI).

Support for decimal and varint data types with SAI indexed columns. (DSP-20794 and DSP-17647)

SAI metrics can be exposed via Metrics Collector and Grafana. DataStax provides prepackaged Grafana dashboards that integrate with DSE Metrics Collector. Use DSE Metrics Collector to export metrics to a monitoring tool like Prometheus, and then visualize the metrics in the Grafana dashboards. See the GitHub-hosted Metrics Reporter resources. Specifically for SAI metrics, refer to Monitor SAI indexes. (DSP-21111)

Support for collection indexing via SAI, and querying CQL collections. See CREATE CUSTOM INDEX and the SAI collection map examples. (DSP-20014)

Fixed a bug where on clusters with small token counts sstableloader could fail. (DB-4463)

None.

None

None

None

Search-related latency metrics will now decay in time like other metrics. Named queries (using query.name parameter) will now have separate latency metrics. New mbean atributes are available for search latency metrics: TotalLatency (us), Min, Max, Mean, StdDev, DurationUnit, MeanRate, OneMinuteRate, FiveMinuteRate, FifteenMinuteRate, RateUnit, 98th, 999th. (DSP-19612)

Fixed a bug where after several months of up time an encrypted index wouldn’t accept more writes unless the core is reloaded. (DSP-21234)

This change allows traversal in to UDTs via the 'properties' and 'values' steps. (DSP-21248)

Fixed type mismatch errors in graph traversals that use simplePath() and as() in presence of search index on edges with composite vertex keys (DSP-21309)

See DataStax Kubernetes Operator for Apache Cassandra (Cass Operator) release notes.

See DataStax Studio 6.8 release notes.

See DataStax Bulk Loader release notes.

See DataStax Apache Kafka Connector release notes.

Apache Solr™ 6.0.1.4.2769

Apache Spark™ 2.4.0.16

Apache TinkerPop™ 3.4.5 with additional production-certified changes

Apache Tomcat® 8.0.53

DSE Java Driver 1.10.0-dse+20200217

Netty 4.1.25.7.dse

Spark JobServer 0.8.0.49

New systemd units for package installs on RHEL and other compatible operating systems. See RedHat systemd configuration. (DSP-7603)

New nodetool import command that lets you import SSTables from one or more directories. See nodetool import. (DB-3253)

New disk-usage guardrail, disk_usage_max_disk_size_in_gb, added in cassandra.yaml. See disk_usage_max_disk_size_in_gb. (DB-4380)

New snapshot_before_dropping_column configuration option added in cassandra.yaml. See snapshot_before_dropping_column.(DB-2690)

Increased the default direct_reads_size_in_mb value in cassandra.yaml. See direct_reads_size_in_mb. (DB-4348)

New entries to jvm.options to assist with capturing thread dumps. (DSP-20778)

If encryption is set on an SSTable, the SAI indexes are encrypted as well. Specifically the entire trie string terms dictionary is encrypted. Also, the user-readable sections of the kdtree are encrypted. See About SAI encryption. (DSP-16939)

Storage-Attached Indexing (SAI) adds support for queries using SAI indexes based on INET columns (IPv4 and IPv6). The queries may use EQ, GT, LT, GE, and LE operators. (DSP-17734)

Storage-Attached Indexing (SAI), a beta feature, adds supports for defining an index based on an individual column that is part of the table’s composite partition key. You can define separate, additional SAI indexes on other individual columns in the same table’s composite partition key. SAI 6.8.x continues to support defining multiple SAI indexes on any non-partition key column in the same table. See Examine SAI column index and query rules. (DSP-20634)

PKCS11 keystore is supported in Advanced workloads. See keystore_type. (DSP-20094)

Python 2.7.x and 3.6+ are supported for cqlsh. (DB-4151)

When working on encrypted SSTables, Cassandra tools such as sstabledump or sstablesplit need to create a remote connection in ReplicatedKeyProvider to a Cassandra node in order to be able to decrypt the sstables. (The sstableloader tool is not affected.) However, if security is enabled on the cluster, the tools won’t be able to decrypt the sstables because they have no way of specifying security parameters for the cluster connection. (DSP-12666)

LDAP servers can now handle multiple, comma separated addresses, with or without a port. If the port is not provided, the ldap_options.server_port parameter is used by default. See LDAP options. (DSP-13086)

Recording a slow CQL query to the log will no longer block the thread. (DSP-20894)

Fixed an issue where after a node replacement procedure, the bootstrap indexing in DSE Search was happening only on one TPC core. (DB-4049)

Fixed an issue to prevent an unbounded number of flushing tasks for memtables that are almost empty. (DB-4376)

Fixed an issue that caused nodes to crash after an upgrade from DSE 5.1 to 6.7.7. (DB-4379)

Fixed the problem with unrepairable digest mismatch between nodes running DSE 5.0.x and DSE 5.1+ (during upgrade) where the query includes a subset of columns. (DB-4399)

Fix an issue that was causing excessive contention during encryption/decryption operations. The fix resulted in an encryption/decryption performance improvement. (DB-4419)

Fixed an issue where during a large test run that reached a density of about 9TB/node, several instances of a TPC WouldBlockException messages, related to NodeSync, appeared in the logs of multiple nodes. (DB-4456)

Fixed a problem with the treatment of zeroes in the type decimal that could cause assertion errors, or not being able to find some rows if their key is 0 written using different precisions, or both. (DB-4472)

Fixed the NullPointerException issue described in CASSANDRA-14200: NPE when dumping sstable with null value for timestamp column. (DB-4512)

An issue that occurred while attempting to create an SSTable index (SASI) in 6.8.0 has been fixed. Previously, attempts to create a SASI index resulted in a NoSuchMethod error. (DSP-20720)

Fixed an issue where the bloom filter false-positive rate calculation did not properly take into account true negatives, such as cases where read operations were attempted on non-existing rows. (DB-3246)

None.

Enhanced LDAP settings to properly handle nested groups so that LDAP enumerates all ancestors of a user’s distinguishedName. (DSP-20107)

Two new LDAP options in dse.yaml (DSP-12612)

While there is no change in default behavior, there is a new render_cql_literals option in dse.yaml under audit logging section, which is false by default. When enabled, bound variables for logged statements will be rendered as CQL literals, which means there will be additional quotation marks and escaping, as well as values of all complex types (collections, tuples, UDTs) will be in human readable format. (DSP-17032)

When DSE tries one authentication scheme and finds that the password is invalid, DSE now tries another scheme, but only if the user has a scheme permission for that other scheme. Example scenarios:

if user has permission to authenticate with scheme A, try scheme A
   if authenticated with A, success, exit
if user has permission to authenticate with scheme B, try scheme B
   if authenticated with B, success, exit
otherwise
   fail, exit

Raised the upper bound limit on DSE LDAP caches. See ldap_options.credentials_validity_in_ms and ldap_options.search_validity_in_seconds. (DSP-21072)

Provided information about how to get tools — such as sstabledump, sstablerepairedset, and sstableloader — working with SSTables that are encrypted with Transparent Data Encryption. See Using tools with TDE-encrypted SSTables. (DSP-20940)

CVE security issues:

Fixed some security vulnerabilities for Solr HTTP REST API when authorization is enabled. Now, users with no appropriate permissions can perform search operations. Resources can be deleted when authorization is enabled, given the correct permissions. (DSP-20749)

Fixed an issue where the audit logging did not capture search queries. (DSP-21058)

Fixed an error condition when DSE failed to get the LDAP roles while refreshing a database schema. (DSP-21075)

In DSEFS, fsck (file system consistency check) throttling is possible via p or -parallelism arguments. In the following example, the command performs fsck by repairing up to 8 files at a time:

dse fs fsck -p 8

Backported the fix from DSP-15762: optimize remove recursive implementation - to lower the tombstone impact on Spark jobs. (DSP-20750)

Fixed excessive number of connections. (DSP-21021)

PKCS11 keystore is supported in Advanced workloads. See keystore_type. (DSP-20094)

Reduced the encrypted search core loading/reloading time. (DSP-20692)

Fixed some security vulnerabilities for Solr HTTP REST API when authorization is enabled. Now, users with no appropriate permissions can perform search operations. Resources can be deleted when authorization is enabled, given the correct permissions. (DSP-20749)

Fixed an issue where a decryption block cache occasionally was not operational (SOLR-14498) (DSP-20987)

Fixed an issue where the audit logging did not capture search queries. (DSP-21058)

Fixed an issue where, after an LCM upgrade, the Solr http API was broken either through curl or the Solr UI. (DSP-21115)

Fixed an issue where after a node replacement procedure, the bootstrap indexing in DSE Search was happening only on one TPC core. (DB-4049)

See DataStax Kubernetes Operator for Apache Cassandra (Cass Operator) release notes.

See DataStax Studio 6.8 release notes.

See DataStax Bulk Loader release notes.

See DataStax Apache Kafka Connector release notes.

Apache Solr™ 6.0.1.4.2746 *

Apache Spark™ 2.4.0.14 *

Apache TinkerPop™ 3.4.5 with additional production-certified changes *

Apache Tomcat® 8.0.53

DSE Java Driver 1.10.0-dse+20200217

Netty 4.1.25.6.dse

Spark JobServer 0.8.0.49 *

A connection pool is created of each LDAP server separately and once the connection is attempted, the best pool is chosen using a simple heuristic. Circuit breaker is used for temporarily disabling those servers which frequently fail to connect. Also, the pool is chosen that has the greatest number of idle connections. Failover parameters are configured through system properties. A new method was added to LDAP mbean to reset LDAP connections - that is, close all connection pools and recreate them.

Added entries to jvm.options to assist with capturing thread dumps. (DSP-20778)

The frequency of range queries performed by lease manager is now configurable via dse.lease.refresh.interval.secondssystem property, JMX, and dsetool command. (DSP-20696)

LDAP servers can now handle multiple, comma-separated addresses, with or without a port. If a port is not provided, the port parameter is used by default so there is no change in configuration for existing users who have LDAP configured. (DSP-13086)

DecayingEstimatedHistogram now decays even when there are no updates, so invalid metric values do not linger. (DSP-20674)

Recording a slow CQL query to the log will no longer block the thread. (DSP-20894)

Added functionality to query role_stats when stats is enabled under role_management_options in dse.yaml. (DB-4283)

Fixed issue of failing to restore backup schema for locally-stored backups. This issue did not affect backups stored on a shared directory or remote blob store. (DB-4297)

Fixed StackOverflowError thrown during read repairs affecting only large clusters or clusters with enabled vnodes. (DB-4350)

When internode_encryption: dc, new nodes fail to bootstrap. (DB-4435)

Increased default direct_reads_size_in_mb value. Previously it was 2M per core + 2M shared. It is now 4M per core + 4M shared. (DB-4348)

The replica side filtering dtests test_update_on_wide_table and test_complementary_update_with_limit_on_static_column_with_not_empty_partitions are more reliable. (DB-4043)

Fixed a problem where continuous paging sessions could leak if the continuous result sets on the driver side were not exhausted or cancelled. Now, when the client connection is terminated, all the corresponding internal continuous paging sessions are cancelled as well. (DB-4313)

Allows DSE non-reserved keywords to be used as user-defined type (UDT) name. (DB-4387)

Fixed bug that caused nodes to crash after an upgrade from 5.1 to 6.7.7. (DB-4379)

New command to asynchronously enable/disable NodeSync on many nodes with nodesync service enable/disable [-n nodes]. (DB-4386)

Fixed read inconsistencies described in CASSANDRA-12126. (DB-3873)

Improved the estimated values of histogram percentiles reported via JMX. In some cases, the percentiles may go slightly up for DSE 6.0+. (DB-4275)

The stream_max_outbound_buffers_in_kb configuration option is now optimally tuned based on the desired stream_throughput_outbound_megabits_per_sec and inter_dc_stream_throughput_outbound_megabits_per_sec; the default value of 0 is recommended. (DB-4063)

Reduced impact of streaming on performance of source nodes. (DB-4372)

Added --disable-history option to cqlsh that disables saving history to disk for current execution. Added history section to cqlshrc which is called with boolean parameter disabled that is set to False by default. (DB-3843)

Security updates:

None.

Security updates:

Internal continuous paging sessions were not closed when LIMIT clause was added in SQL query, which caused sessions leak and inability to close the Spark application gracefully because the Java driver waited indefinitely for orphaned sessions to finish. (DSP-19804)

Security updates:

Fixed search queries in core graph for complex types such as date and time. (DSP-20685)

Prior to this ticket, running any ALTER TABLE ... WITH table_properties CQL statement would strip the an existing table’s vertex label or edge label. Now those should be preserved. (DSP-20901)

Reduced the encrypted search core loading and reloading time. (DSP-20692)

Fixed the suggest request handler to require select permission. Previously, suggest request handler returned forbidden response when authorization was on, regardless of the user permissions. (DSP-20697)

Security updates:

See DataStax Studio 6.8 release notes.

See DataStax Bulk Loader release notes.

See DataStax Apache Kafka Connector release notes.

Catch non-IOException in FileUtils.close to make sure that all resources are closed (CASSANDRA-15225)

Nodetool import row cache invalidation races with adding SSTables to tracker (CASSANDRA-14529)

Let nodetool import take a list of directories (CASSANDRA-14442)

nodetool import cleanup and improvements (CASSANDRA-14417)

Add ability to load new SSTables from a separate directory (CASSANDRA-6719)

Add a few options to nodetool verify (CASSANDRA-14201)

Make all DDL statements idempotent and not dependent on global state (CASSANDRA-13426)

Always close RT markers returned by ReadCommand#executeLocally() (CASSANDRA-14515)

BloomFilter serialization format should not change byte ordering (CASSANDRA-9067)

Remove unused on-heap BloomFilter implementation (CASSANDRA-14152)

Add a virtual table to expose settings (CASSANDRA-14573)

Add a virtual table to expose caches (CASSANDRA-14538)

Fix up chunk cache handling of metrics (CASSANDRA-14628)

Add a virtual table to expose active client connections (CASSANDRA-14458)

Clean up and refactor client metrics (CASSANDRA-14524)

NodeTool clientstats should show SSL Cipher (CASSANDRA-14322)

Add ability to specify driver name and version (CASSANDRA-14275)

Add nodetool clientlist (CASSANDRA-13665)

Pad uncompressed chunks when they would be interpreted as compressed (CASSANDRA-14892)

Bind to correct local address in 4.0 streaming (CASSANDRA-14362)

Set broadcast address in internode messaging handshake (CASSANDRA-14579)

Internode messaging handshake sends wrong messaging version number (CASSANDRA-14540)

Use Netty for streaming (CASSANDRA-12229)

Use Netty for internode messaging (CASSANDRA-8457)

Correct and clarify SSLFactory.getSslContext method and call sites (CASSANDRA-14314)

Properly close StreamCompressionInputStream to release any ByteBuf (CASSANDRA-13906)

Correctly close Netty channels when a stream session ends (CASSANDRA-13905)

Fix buffer length comparison when decompressing in Netty-based streaming (CASSANDRA-13899)

Race condition when closing stream sessions (CASSANDRA-13852)

Make monotonic read / read repair configurable (CASSANDRA-14635)

Improve read repair blocking behavior (CASSANDRA-10726)

Add coordinator write metric per CF (CASSANDRA-14232)

Make PartitionUpdate and Mutation immutable (CASSANDRA-13867)

Disable old native protocol versions on demand (CASSANDRA-14659)

Refactor CompactionStrategyManager (CASSANDRA-14621)

Extend IAuthenticator to accept peer SSL certificates (CASSANDRA-14652)

For LCS, single SSTable up-level is handled inefficiently (CASSANDRA-12526)

Fix setting min/max compaction threshold with LCS (CASSANDRA-14388)

Support light-weight transactions in cassandra-stress (CASSANDRA-13529)

Add a virtual table to expose all running sstable tasks (CASSANDRA-14457)

Implement virtual keyspace interface (CASSANDRA-7622)

cassandra-stress throws NPE if insert section isn’t specified in user profile (CASSSANDRA-14426)

nodetool listsnapshots is missing local system keyspace snapshots (CASSANDRA-14381)

CVE-2017-5929 Security vulnerability and redefine default log rotation policy (CASSANDRA-14183)

Fix sstablemetadata date string for minLocalDeletionTime (CASSANDRA-14132)

Make sub-range selection for non-frozen collections return null instead of empty (CASSANDRA-14182)

Fix cassandra-stress startup failure (CASSANDRA-14106)

Fix trivial log format error (CASSANDRA-14015)

Allow sstabledump to do a json object per partition (CASSANDRA-13848)

Remove unused and deprecated methods from AbstractCompactionStrategy (CASSANDRA-14081)

Fix Distribution.average in cassandra-stress (CASSANDRA-14090)

Presize collections (CASSANDRA-13760)

Add GroupCommitLogService (CASSANDRA-13530)

Parallelize initial materialized view build (CASSANDRA-12245)

Fix flaky SecondaryIndexManagerTest.assert[Not]MarkedAsBuilt (CASSANDRA-13965)

Make LWTs send resultset metadata on every request (CASSANDRA-13992)

Fix flaky indexWithFailedInitializationIsNotQueryableAfterPartialRebuild (CASSANDRA-13963)

Introduce leaf-only iterator (CASSANDRA-9988)

Allow only one concurrent call to StatusLogger (CASSANDRA-12182)

Refactoring to specialised functional interfaces (CASSANDRA-13982)

Speculative retry should allow more friendly params (CASSANDRA-13876)

Throw exception if we send/receive repair messages to incompatible nodes (CASSANDRA-13944)

Replace usages of MessageDigest with Guava’s Hasher (CASSANDRA-13291)

Add nodetool cmd to print hinted handoff window (CASSANDRA-13728)

Fix some alerts raised by static analysis (CASSANDRA-13799)

Checksum sstable metadata (CASSANDRA-13321, CASSANDRA-13593)

Add result set metadata to prepared statement MD5 hash calculation (CASSANDRA-10786)

Add incremental repair support for --hosts, --force, and subrange repair (CASSANDRA-13818)

Refactor GcCompactionTest to avoid boxing (CASSANDRA-13941)

Expose recent histograms in JmxHistograms (CASSANDRA-13642)

Add SERIAL and LOCAL_SERIAL support for cassandra-stress (CASSANDRA-13925)

LCS needlessly checks for L0 STCS candidates multiple times (CASSANDRA-12961)

Update lz4 to 1.4.0 (CASSANDRA-13741)

Throttle base partitions during MV repair streaming to prevent OOM (CASSANDRA-13299)

Improve short read protection performance (CASSANDRA-13794)

Fix AssertionError in short read protection (CASSANDRA-13747)

Use compaction threshold for STCS in L0 (CASSANDRA-13861)

Fix problem with min_compress_ratio: 1 and disallow ratio < 1 (CASSANDRA-13703)

Add extra information to SASI timeout exception (CASSANDRA-13677)

Rework CompactionStrategyManager.getScanners synchronization (CASSANDRA-13786)

Add additional unit tests for batch behavior, TTLs, Timestamps (CASSANDRA-13846)

Add keyspace and table name in schema validation exception (CASSANDRA-13845)

Emit metrics whenever we hit tombstone failures and warn thresholds (CASSANDRA-13771)

Allow changing log levels via nodetool for related classes (CASSANDRA-12696)

Add stress profile yaml with LWT (CASSANDRA-7960)

Reduce memory copies and object creations when acting on ByteBufs (CASSANDRA-13789)

simplify mx4j configuration (Cassandra-13578)

Fix trigger example on 4.0 (CASSANDRA-13796)

force minimum timeout value (CASSANDRA-9375)

Add bytes repaired/unrepaired to nodetool tablestats (CASSANDRA-13774)

Don’t delete incremental repair sessions if they still have sstables (CASSANDRA-13758)

Fix pending repair manager index out of bounds check (CASSANDRA-13769)

Don’t use RangeFetchMapCalculator when RF=1 (CASSANDRA-13576)

Don’t optimise trivial ranges in RangeFetchMapCalculator (CASSANDRA-13664)

Use an ExecutorService for repair commands instead of new Thread(..).start() (CASSANDRA-13594)

Fix race / ref leak in anticompaction (CASSANDRA-13688)

Fix race / ref leak in PendingRepairManager (CASSANDRA-13751)

Enable ppc64le runtime as unsupported architecture (CASSANDRA-13615)

Improve sstablemetadata output (CASSANDRA-11483)

Support for migrating legacy users to roles has been dropped (CASSANDRA-13371)

Introduce error metrics for repair (CASSANDRA-13387)

Refactoring to primitive functional interfaces in AuthCache (CASSANDRA-13732)

Update metrics to 3.1.5 (CASSANDRA-13648)

batch_size_warn_threshold_in_kb can now be set at runtime (CASSANDRA-13699)

Avoid always rebuilding secondary indexes at startup (CASSANDRA-13725)

Upgrade JMH from 1.13 to 1.19 (CASSANDRA-13727)

Upgrade SLF4J from 1.7.7 to 1.7.25 (CASSANDRA-12996)

Default for start_native_transport now true if not set in config (CASSANDRA-13656)

Don’t add localhost to the graph when calculating where to stream from (CASSANDRA-13583)

Allow skipping equality-restricted clustering columns in ORDER BY clause (CASSANDRA-10271)

Use common nowInSec for validation compactions (CASSANDRA-13671)

Improve handling of IR prepare failures (CASSANDRA-13672)

Send IR coordinator messages synchronously (CASSANDRA-13673)

Flush system.repair table before IR finalize promise (CASSANDRA-13660)

Fix column filter creation for wildcard queries (CASSANDRA-13650)

Add 'nodetool getbatchlogreplaythrottle' and 'nodetool setbatchlogreplaythrottle' (CASSANDRA-13614)

fix race condition in PendingRepairManager (CASSANDRA-13659)

Allow noop incremental repair state transitions (CASSANDRA-13658)

Run repair with down replicas (CASSANDRA-10446)

Added started & completed repair metrics (CASSANDRA-13598)

Added started & completed repair metrics (CASSANDRA-13598)

Improve secondary index (re)build failure and concurrency handling (CASSANDRA-10130)

Improve calculation of available disk space for compaction (CASSANDRA-13068)

Change the accessibility of RowCacheSerializer for third party row cache plugins (CASSANDRA-13579)

Allow sub-range repairs for a preview of repaired data (CASSANDRA-13570)

NPE in IR cleanup when columnfamily has no sstables (CASSANDRA-13585)

Fix Randomness of stress values (CASSANDRA-12744)

Allow selecting Map values and Set elements (CASSANDRA-7396)

Fast and garbage-free Streaming Histogram (CASSANDRA-13444)

Update repairTime for keyspaces on completion (CASSANDRA-13539)

Add configurable upper bound for validation executor threads (CASSANDRA-13521)

Bring back maxHintTTL propery (CASSANDRA-12982)

Add testing guidelines (CASSANDRA-13497)

Add more repair metrics (CASSANDRA-13531)

RangeStreamer should be smarter when picking endpoints for streaming (CASSANDRA-4650)

Avoid rewrapping an exception thrown for cache load functions (CASSANDRA-13367)

Log time elapsed for each incremental repair phase (CASSANDRA-13498)

Add multiple table operation support to cassandra-stress (CASSANDRA-8780)

Cassandra is now relying on the JVM options to properly shutdown on OutOfMemoryError. By default it will rely on the OnOutOfMemoryError option as the ExitOnOutOfMemoryError and CrashOnOutOfMemoryError options are not supported by the older 1.7 and 1.8 JVMs. A warning will be logged at startup if none of those JVM options are used. See CASSANDRA-13006 for more details.

Creating Materialized View with filtering on non-primary-key base column (added in CASSANDRA-10368) is disabled, because the liveness of view row is depending on multiple filtered base non-key columns and base non-key column used in view primary-key. This semantic cannot be supported without storage format change, see CASSANDRA-13826. For append-only use case, you may still use this feature with a startup flag: "-Dcassandra.mv.allow_filtering_nonkey_columns_unsafe=true"

The NativeAccessMBean isAvailable method will only return true if the native library has been successfully linked. Previously it was returning true if JNA could be found but was not taking into account link failures.

Primary ranges in the system.size_estimates table are now based on the keyspace replication settings and adjacent ranges are no longer merged (CASSANDRA-9639).

In 2.1, the default for otc_coalescing_strategy was 'DISABLED'. In 2.2 and 3.0, it was changed to 'TIMEHORIZON', but that value was shown to be a performance regression. The default for 3.11.0 and newer has been reverted to 'DISABLED'. Users upgrading from Cassandra 2.2 or 3.0 should be aware that the default has changed.

The StorageHook interface has been modified to allow to retrieve read information from SSTableReader (CASSANDRA-13120).

Materialized Views for upgrades from DSE 5.1.1 or 5.1.2 or any version DSE 5.0.10 or later:

Runtime modification of concurrent_compactors is now available via nodetool concurrent_compactors.

Support for the assignment operators +=/-= has been added for update queries.

An Index implementation may now provide a task which runs prior to joining the ring. See CASSANDRA-12039

Filtering on partition key columns is now also supported for queries without secondary indexes.

A slow query log has been added: slow queries will be logged at DEBUG level. For more details refer to CASSANDRA-12403 and slow_query_log_timeout_in_ms in cassandra.yaml.

Support for GROUP BY queries has been added.

A new compaction-stress tool has been added to test the throughput of compaction for any cassandra-stress user schema. see compaction-stress help for how to use.

Prepared statements are now persisted in the table prepared_statements in the system keyspace. Upon startup, this table is used to preload all previously prepared statements - i.e. in many cases clients do not need to re-prepare statements against restarted nodes.

cqlsh can now connect to older Cassandra versions by downgrading the native protocol version. Please note that this is currently not part of our release testing and, as a consequence, it is not guaranteed to work in all cases. See CASSANDRA-12150 for more details.

Snapshots that are automatically taken before a table is dropped or truncated will have a "dropped" or "truncated" prefix on their snapshot tag name.

Metrics are exposed for successful and failed authentication attempts. These can be located using the object names org.apache.cassandra.metrics:type=Client,name=AuthSuccess and org.apache.cassandra.metrics:type=Client,name=AuthFailure respectively.

Add support to "unset" JSON fields in prepared statements by specifying DEFAULT UNSET. See CASSANDRA-11424 for details

Allow TTL with null value on insert and update. It will be treated as equivalent to inserting a 0.

Removed outboundBindAny configuration property. See CASSANDRA-12673 for details.

Support for alter types of already defined tables and of UDTs fields has been disabled. If it is necessary to return a different type, please use casting instead. See CASSANDRA-12443 for more details.

Specifying the default_time_to_live option when creating or altering a materialized view was erroneously accepted (and ignored). It is now properly rejected.

Only Java and JavaScript are now supported UDF languages. The sandbox in 3.0 already prevented the use of script languages except Java and JavaScript.

Compaction now correctly drops sstables out of CompactionTask when there isn’t enough disk space to perform the full compaction. This should reduce pending compaction tasks on systems with little remaining disk space.

Request timeouts in cassandra.yaml (read_request_timeout_in_ms, etc) now apply to the "full" request time on the coordinator. Previously, they only covered the time from when the coordinator sent a message to a replica until the time that the replica responded. Additionally, the previous behavior was to reset the timeout when performing a read repair, making a second read to fix a short read, and when subranges were read as part of a range scan or secondary index query. In 3.10 and higher, the timeout is no longer reset for these "subqueries". The entire request must complete within the specified timeout. As a consequence, your timeouts may need to be adjusted to account for this. See CASSANDRA-12256 for more details.

Logs written to stdout are now consistent with logs written to files. Time is now local (it was UTC on the console and local in files). Date, thread, file and line info where added to stdout. (see CASSANDRA-12004)

The 'clientutil' jar, which has been somewhat broken on the 3.x branch, is not longer provided. The features provided by that jar are provided by any good java driver and we advise relying on drivers rather on that jar, but if you need that jar for backward compatiblity until you do so, you should use the version provided on previous Cassandra branch, like the 3.0 branch (by design, the functionality provided by that jar are stable accross versions so using the 3.0 jar for a client connecting to 3.x should work without issues).

(Tools development) DatabaseDescriptor no longer implicitly startups components/services like commit log replay. This may break existing 3rd party tools and clients. In order to startup a standalone tool or client application, use the DatabaseDescriptor.toolInitialization() or DatabaseDescriptor.clientInitialization() methods. Tool initialization sets up partitioner, snitch, encryption context. Client initialization just applies the configuration but does not setup anything. Instead of using Config.setClientMode() or Config.isClientMode(), which are deprecated now, use one of the appropiate new methods in DatabaseDescriptor.

Application layer keep-alives were added to the streaming protocol to prevent idle incoming connections from timing out and failing the stream session (CASSANDRA-11839). This effectively deprecates the streaming_socket_timeout_in_ms property in favor of streaming_keep_alive_period_in_secs. See cassandra.yaml for more details about this property.

Duration literals support the ISO 8601 format. By consequence, identifiers matching that format (e.g P2Y or P1MT6H) will not be supported anymore (CASSANDRA-11873).

Shared pool threads are now named according to the stage they are executing tasks for. Thread names mentioned in traced queries change accordingly.

A new option has been added to cassandra-stress "-rate fixed={number}/s" that forces a scheduled rate of operations/sec over time. Using this, stress can accurately account for coordinated ommission from the stress process.

The cassandra-stress "-rate limit=" option has been renamed to "-rate throttle="

hdr histograms have been added to stress runs, it’s output can be saved to disk using: "-log hdrfile=" option. This histogram includes response/service/wait times when used with the fixed or throttle rate options. The histogram file can be plotted on http://hdrhistogram.github.io/HdrHistogram/plotFiles.html

TimeWindowCompactionStrategy has been added. This has proven to be a better approach to time series compaction and new tables should use this instead of DTCS. See CASSANDRA-9666 for details.

DateTieredCompactionStrategy has been deprecated - new tables should use TimeWindowCompactionStrategy. Note that migrating an existing DTCS-table to TWCS might cause increased compaction load for a while after the migration so make sure you run tests before migrating. Read CASSANDRA-9666 for background on this.

Change-Data-Capture is now available. See cassandra.yaml and for cdc-specific flags and a brief explanation of on-disk locations for archived data in CommitLog form. This can be enabled via ALTER TABLE …​ WITH cdc=true. Upon flush, CommitLogSegments containing data for CDC-enabled tables are moved to the data/cdc_raw directory until removed by the user and writes to CDC-enabled tables will be rejected with a WriteTimeoutException once cdc_total_space_in_mb is reached between unflushed CommitLogSegments and cdc_raw.

The ReversedType behaviour has been corrected for clustering columns of BYTES type containing empty value. Scrub should be run on the existing SSTables containing a descending clustering column of BYTES type to correct their ordering. See CASSANDRA-12127 for more details.

Ec2MultiRegionSnitch will no longer automatically set broadcast_rpc_address to the public instance IP if this property is defined on cassandra.yaml.

The name "json" and "distinct" are not valid anymore a user-defined function names (they are still valid as column name however). In the unlikely case where you had defined functions with such names, you will need to recreate those under a different name, change your code to use the new names and drop the old versions, and this before upgrade (see CASSANDRA-10783 for more details).

A maximum size for SSTables values has been introduced, to prevent out of memory exceptions when reading corrupt SSTables. This maximum size can be set via max_value_size_in_mb in cassandra.yaml. The default is 256MB, which matches the default value of native_transport_max_frame_size_in_mb. SSTables will be considered corrupt if they contain values whose size exceeds this limit. See CASSANDRA-9530 for more details.

JMX connections can now use the same auth mechanisms as CQL clients. New options in cassandra-env.(sh|ps1) enable JMX authentication and authorization to be delegated to the IAuthenticator and IAuthorizer configured in cassandra.yaml. The default settings still only expose JMX locally, and use the JVM’s own security mechanisms when remote connections are permitted. For more details on how to enable the new options, see the comments in cassandra-env.sh. A new class of IResource, JMXResource, is provided for the purposes of GRANT/REVOKE via CQL. See CASSANDRA-10091 for more details. Also, directly setting JMX remote port via the com.sun.management.jmxremote.port system property at startup is deprecated. See CASSANDRA-11725 for more details.

JSON timestamps are now in UTC and contain the timezone information, see CASSANDRA-11137 for more details.

Collision checks are performed when joining the token ring, regardless of whether the node should bootstrap. Additionally, replace_address can legitimately be used without bootstrapping to help with recovery of nodes with partially failed disks. See CASSANDRA-10134 for more details.

Key cache will only hold indexed entries up to the size configured by column_index_cache_size_in_kb in cassandra.yaml in memory. Larger indexed entries will never go into memory. See CASSANDRA-11206 for more details.

For tables having a default_time_to_live specifying a TTL of 0 will remove the TTL from the inserted or updated values.

Startup is now aborted if corrupted transaction log files are found. The details of the affected log files are now logged, allowing the operator to decide how to resolve the situation.

Filtering expressions are made more pluggable and can be added programatically via a QueryHandler implementation. See CASSANDRA-11295 for more details.

Internal authentication now supports caching of encrypted credentials. Reference cassandra.yaml:credentials_validity_in_ms

Remote configuration of auth caches via JMX can be disabled using the the system property cassandra.disable_auth_caches_remote_configuration

sstabledump tool is added to be 3.0 version of former sstable2json. The tool only supports v3.0+ SSTables. See tool’s help for more detail.

The mbean interfaces org.apache.cassandra.auth.PermissionsCacheMBean and org.apache.cassandra.auth.RolesCacheMBean are deprecated in favor of org.apache.cassandra.auth.AuthCacheMBean. This generalized interface is common across all caches in the auth subsystem. The specific mbean interfaces for each individual cache will be removed in a subsequent major version.

We now make sure that a token does not exist in several data directories. This means that we run one compaction strategy per data_file_directory and we use one thread per directory to flush. Use nodetool relocatesstables to make sure your tokens are in the correct place, or just wait and compaction will handle it. See CASSANDRA-6696 for more details.

bound maximum in-flight commit log replay mutation bytes to 64 megabytes tunable via cassandra.commitlog_max_outstanding_replay_bytes

Support for type casting has been added to the selection clause.

Hinted handoff now supports compression. Reference cassandra.yaml:hints_compression.

The Thrift API is deprecated and will be removed in Cassandra 4.0.

The compression ratio metrics computation has been modified to be more accurate.

Running Cassandra as root is prevented by default.

JVM options are moved from cassandra-env.(sh|ps1) to jvm.options.

The return value of SelectStatement::getLimit as been changed from DataLimits to int.

Custom index implementation should be aware that the method Indexer::indexes() has been removed as its contract was misleading and all custom implementation should have almost surely returned true inconditionally for that method.

GC logging is now enabled by default (you can disable it in the jvm.options file if you prefer).

Apache Solr™ 6.0.1.4.2718

Apache Spark™ 2.4.0.13

Apache TinkerPop™ 3.4.5 with additional production-certified changes

Apache Tomcat® 8.0.53

DSE Java Driver 1.10.0-dse+20200217

Netty 4.1.25.6.dse

Spark Jobserver 0.8.0.49 (DSE custom version)

DSE 6.8 is based on Apache Cassandra® 3.11 with additional production-certified changes.

Beta version of Storage-Attached Indexing (SAI), which provides traditional relational database indexing capabilities for DSE, without complexity or operational challenges. Includes the revised CREATE CUSTOM INDEX command with a new USING 'StorageAttachedIndex' clause. See the SAI topics and the CREATE CUSTOM INDEX command.

Beta version of the DSE Backup and Restore Service, which enables cluster-wide backup and restore operations. Intended for users of DataStax Kubernetes Operator for Apache Cassandra® (Cassandra Operator).

Performance improvement: Better memtable memory usage. (DB-2164)

DataStax recommends enabling NodeSync on base tables with materialized views. See Materialized views maintenance guidelines. (DB-2831)

Simplified maintenance of materialized views by requiring periodic repair of only base tables. Updates on columns indexed in materialized views might see a performance regression of up to 20% in throughput and latency. (DB-2816)

New client requests rate and internode message rate usage statistics are output from nodetool tpstats. (DB-3619)

Fixed a rare race during streaming that could have resulted in aborted SSTable files being left on disk. (DB-2733)

Compaction performance improvement with new cassandra.yaml pick_level_on_streaming option. (DB-1658)

Incremental NodeSync service on newly created tables is now on by default. (DB-3934)

Introduce new TRUNCATE permission. (DB-74)

Create new settings system view by merging CASSANDRA-14573. Allows select properties to be modified. (DB-37)

In cassandra.yaml, memtable_flush_writers now defaults to 8 and memtable_cleanup_threshold now defaults to max(0.2, 1 / (memtable_flush_writers + 1)) to establish a lower bound. PendingTasks metric attached to all JMXEnabledThreadPoolExecutor instances, including MemtableFlushWriter, now excludes currently executing tasks. (DB-2871)

Fix error that occurs when initialCapacity overflows during continuous paging. (DB-2852)

dsetool ring shows in-progress search index building during bootstrap. (DSP-15281)

Improved output of nodetool tablestats to accurately report when no data is available. (DB-2854)

New cqlsh command line option to set consistency-level to serial. (DB-4225)

Remove the serialization header partition/clustering key validation for secondary-index sstables and changes the validation to consider the element type. (DB-4111)

Keyspace and table names now allow 222 characters. (DB-2128)

Fix viewbuildstatus error that returns incorrect value. (DB-2397)

nodetool stop has anticompaction option. (DB-3821)

New ALL TABLES permission allows permissions on all of the tables of a keyspace, but not the keyspace itself. (DB-3823)

Log messages for leak detection now specify if the leaked resource has negative consequences. (DB-3827)

Fix potentially incorrect dropped messages in case of time drifts. (DB-3891)

Change in validation method for initial_token and num_tokens for cassandra.yaml. (DB-1618)

Fix a bug to correctly estimate the size of the on-heap memtable metadata. (DB-2086)

Starting from DSE 6.8.0, if nodes need to communicate via a different interface than the one configured as the listen_address, you must configure the additional interface as the broadcast_address. If no routing exists between the two interfaces, you need to eventually set listen_on_broadcast_address: true. (DB-4142)

Fix bug to allow LIST ROLES and LIST USERS to work with system-keyspace-filtering enabled. (DB-4221)

nodetool compactionstats now accurately shows the pending tasks for TimeWindowCompactionStrategy (TWCS). (DB-3495)

DROP KEYSPACE now waits for interrupted compactions to finish before dropping the keyspace. (DB-3575)

During the execution of CQL queries, check guardrails are ignored if the user is a superuser or for internal system queries. (DB-3654)

Replaced stream_entire_sstables for Cassandra with zerocopy_streaming_enabled for DataStax Enterprise, along with the related options. (DB-3832)

Improved handling of SSTable min and max clustering. (DB-3728)

For cassandra.yaml configuration, only TLS is allowed for the protocol option for client_encryption_options and server_encryption_options for security. (DB-2786)

New counter metrics for submitted traversals and throughput are under com.datastax.bdp.metrics.graph. (DSP-17009)

Improved performance of dsetool, dse client-tool, dse fs, and nodetool commands. (DSP-17586)

DataStax Bulk Loader is not included with DataStax Enterprise installations, but can be installed separately. (DSP-19469)

The application_name and application_version attributes set by the Python driver flow through to driver event reporting and Insights. cqlsh sets these attributes to appropriate values so connected clients appear neatly in Insights and other reporting. (DSP-20119)

Incremental Nodesync is disabled when an ordering partitioner configured. (DB-4024)

Added hostname_verification to ldap_options in dse.yaml. (DSP-20302)

Add the ability to add the reason for re-indexing to the dsetool core_indexing_status command. (DSP-20264)

Make the search reference visible in the error message for LDAP connections. (DSP-20578)

The jvm.options file is now named jvm-server.options. (DSP-20769)

Security updates:

Streaming throughput (bootstrapping, decommissioning nodes) may be slower on networks with high latency. (DB-4041)

Certain workloads may cause higher context switches and increased latencies. (DSP-20499)

-Ddse.tpc.work_stealing_max_unparks=1

NoSuchMethod error returned when creating SASI index. (DSP-20720)

Improved reliability of no-space-left-on-device detection. The DSEFS min_free_space default value in dse.yaml is reduced from 5 GB to 256 MB. (DSP-16873)

Apache Spark™ 2.4 runs with Scala 2.11.12 by default. Upgrade the compile time dependencies for structured streaming and other experimental Spark features. Even though most Spark jobs from earlier Spark 2.x builds can run on Spark 2.4 without recompiling, DataStax recommends that you recompile your applications against Spark 2.4 to guarantee compatibility. (DSP-17823)

Add the ability to set time-to-live (TTL) and WriteTime in DseGraphFrames and Spark DataFrames. (DSP-17044)

Bring-Your-Own-Spark (BYOS) builds include dependencies for Joda and Commons-Configuration. (DSP-20512)

During Spark Application startup, Exception: java.lang.ExceptionInInitializerError thrown from the UncaughtExceptionHandler in thread "main" was logged, sometimes instead of a meaningful error. (DSP-20474)

New spark.cassandra.query.consistency.level parameter sets the default consistency level for sessions accessed by Spark Connector. The default consistency level for HiveMetaStore is LOCAL_QUORUM. (DSP-19982)

Changes to IN clauses. (DSP-15203)

Running Spark applications with large number of partitions creates many tombstones and may cause tombstone warnings or in extreme cases a job failure. DSE 6.8 reduces the number of tombstones created during Spark job commit and improves performance of some Spark jobs up to 60%. (DSP-15762)

When creating a file through WebHDFS API, DSEFS does not verify WX permissions of parent’s parent when and the parent exists. (DSP-20355)

Allow DSEFS to use mixed case keyspaces to connect directly to the dsefs keyspace. (DSP-20354)

DSEFS node identifiers are now the same as DSE node identifiers. The NODE_ID file in the dsefs working directory is no longer needed. (DSP-18009)

Add support for multiple contact points for DSEFS implementation of the Hadoop FileSystem. Provides FileSystem URI with dsefs://host0[:port][,host1[:port]]/. ( DSP-19704)

DSEFS local file system implementation now returns alphabetically sorted directories and files when using wildcards and listing command. (DSP-20057)

DSEFS now stores information about data usage in the local storage directory instead of Cassandra. This change improves reliability if some nodes are down. (DSP-15349)

Improve reliability of DSEFS internode connections. Fix error for missing session key when cluster nodes were down. (DSP-15347)

Improved DSEFS node health reported by dsefs df for consistency with dsetool status and nodetool status. (DSP-15346)

Traversal length is hard-coded to 90 steps and traversals with more steps fail with an error message to split it into multiple smaller traversals. (DSP-17657)

Enhanced Graph OLAP Spark configuration. (DSP-17832)

DataStax Graph (core) changes:

Track latencies for unaliased, global, and per-graph traversals. (DSP-16455)

Upgraded dependency for Graph prototyping to TinkerPop 3.4.0. (DSP-16452)

The AndStep, OrStep, DedupGlobalStep, RangeGlobalStep, NotStep, SelectStep, SelectOneStep, OrderGlobalStep, and WherePredicateStep steps are valid when determining whether to route OLAP traversals to DseGraphFrames. (DSP-16233)

Allow range queries using part of the Custom Vertex ID. This change requires the partition key components to be specified before the range query may be specified on the clustering key components. (DSP-12501)

Add read/write support for TTL and WriteTime into DataGraphFrames (DGF). (DSP-19304)

Expose configuration and metrics for Gremlin query cache. (DSP-20240)

Change classic Graph query so vertices are read from _p tables in Cassandra using SELECT ... WHERE <vertex primary key columns> statement. The search predicate is applied in memory. (DSP-20230)

Update TinkerPop version bump, which changes the following settings if SSL configuration options are used for Gremlin Console and TinkerPop drivers:

Server could get slow or unresponsive if lots of long-running traversals are canceled. (DSP-20425)

Unbounded facet searches are no longer allowed. (DSP-18693)

The dsetool stop_core_reindex command now mentions the node in the output message. (DSP-17090)

Legacy Solr join queries are no longer valid. The to, from, and force parameters are invalid. Joins can no longer be performed on non-partition key columns or on different keyspaces. (DSP-17431)

The dsetool core_indexing_status command now mentions the indexing reason in the output message.

The recommendation to enable live indexing on only one search core per cluster was too conservative. See Tuning search for maximum indexing throughput and Capacity planning for DSE Search. Be sure to follow the DataStax recommendations for your environment. (DSP-17939)

DSE Management API is available for enhancing operation with Kubernetes. (DSP-18785)

Improved real-time search to fix a docValues bug. (DSP-20300)

Passing TextField Solr fields with docValues to facet.field, facet.pivot, group.field, and sort (including native CQL Solr queries that use a TextField with docValues with ORDER BY) is now illegal and will fail the query in question. (DSP-18238)

Improved guidance with warnings when index rebuild is required for ALTER SEARCH INDEX, RELOAD SEARCH INDEX, and dsetool reload_core commands. (DSP-19347)

Replicas with non-queryable indexes will be skipped by the coordinator node to improve availability for index read. During nodetool rebuild_index, Storage-Attached Indexing (SAI) will be marked as non-queryable until the index build finishes, while 2i will remain queryable. (DSP-19543) For related information, see What is SAI?.

Error messages related to Solr errors contain better descriptions of the root cause. (DSP-13792)

Mixed workloads with very wide partitions could see diminished performance. (DSP-20386)

-Dnetty.eventloop.tasks_processing_time_limit_ms=100

See DataStax Studio 6.8 release notes.

See DataStax Bulk Loader release notes.

See DataStax Apache Kafka Connector release notes.

Catch non-IOException in FileUtils.close to make sure that all resources are closed (CASSANDRA-15225)

Nodetool import row cache invalidation races with adding SSTables to tracker (CASSANDRA-14529)

Let nodetool import take a list of directories (CASSANDRA-14442)

nodetool import cleanup and improvements (CASSANDRA-14417)

Add ability to load new SSTables from a separate directory (CASSANDRA-6719)

Add a few options to nodetool verify (CASSANDRA-14201)

Make all DDL statements idempotent and not dependent on global state (CASSANDRA-13426)

Always close RT markers returned by ReadCommand#executeLocally() (CASSANDRA-14515)

BloomFilter serialization format should not change byte ordering (CASSANDRA-9067)

Remove unused on-heap BloomFilter implementation (CASSANDRA-14152)

Add a virtual table to expose settings (CASSANDRA-14573)

Add a virtual table to expose caches (CASSANDRA-14538)

Fix up chunk cache handling of metrics (CASSANDRA-14628)

Add a virtual table to expose active client connections (CASSANDRA-14458)

Clean up and refactor client metrics (CASSANDRA-14524)

NodeTool clientstats should show SSL Cipher (CASSANDRA-14322)

Add ability to specify driver name and version (CASSANDRA-14275)

Add nodetool clientlist (CASSANDRA-13665)

Pad uncompressed chunks when they would be interpreted as compressed (CASSANDRA-14892)

Bind to correct local address in 4.0 streaming (CASSANDRA-14362)

Set broadcast address in internode messaging handshake (CASSANDRA-14579)

Internode messaging handshake sends wrong messaging version number (CASSANDRA-14540)

Use Netty for streaming (CASSANDRA-12229)

Use Netty for internode messaging (CASSANDRA-8457)

Correct and clarify SSLFactory.getSslContext method and call sites (CASSANDRA-14314)

Properly close StreamCompressionInputStream to release any ByteBuf (CASSANDRA-13906)

Correctly close Netty channels when a stream session ends (CASSANDRA-13905)

Fix buffer length comparison when decompressing in Netty-based streaming (CASSANDRA-13899)

Race condition when closing stream sessions (CASSANDRA-13852)

Make monotonic read / read repair configurable (CASSANDRA-14635)

Improve read repair blocking behavior (CASSANDRA-10726)

Add coordinator write metric per CF (CASSANDRA-14232)

Make PartitionUpdate and Mutation immutable (CASSANDRA-13867)

Disable old native protocol versions on demand (CASSANDRA-14659)

Refactor CompactionStrategyManager (CASSANDRA-14621)

Extend IAuthenticator to accept peer SSL certificates (CASSANDRA-14652)

For LCS, single SSTable up-level is handled inefficiently (CASSANDRA-12526)

Fix setting min/max compaction threshold with LCS (CASSANDRA-14388)

Support light-weight transactions in cassandra-stress (CASSANDRA-13529)

Add a virtual table to expose all running sstable tasks (CASSANDRA-14457)

Implement virtual keyspace interface (CASSANDRA-7622)

cassandra-stress throws NPE if insert section isn’t specified in user profile (CASSSANDRA-14426)

nodetool listsnapshots is missing local system keyspace snapshots (CASSANDRA-14381)

CVE-2017-5929 Security vulnerability and redefine default log rotation policy (CASSANDRA-14183)

Fix sstablemetadata date string for minLocalDeletionTime (CASSANDRA-14132)

Make sub-range selection for non-frozen collections return null instead of empty (CASSANDRA-14182)

Fix cassandra-stress startup failure (CASSANDRA-14106)

Fix trivial log format error (CASSANDRA-14015)

Allow sstabledump to do a json object per partition (CASSANDRA-13848)

Remove unused and deprecated methods from AbstractCompactionStrategy (CASSANDRA-14081)

Fix Distribution.average in cassandra-stress (CASSANDRA-14090)

Presize collections (CASSANDRA-13760)

Add GroupCommitLogService (CASSANDRA-13530)

Parallelize initial materialized view build (CASSANDRA-12245)

Fix flaky SecondaryIndexManagerTest.assert[Not]MarkedAsBuilt (CASSANDRA-13965)

Make LWTs send resultset metadata on every request (CASSANDRA-13992)

Fix flaky indexWithFailedInitializationIsNotQueryableAfterPartialRebuild (CASSANDRA-13963)

Introduce leaf-only iterator (CASSANDRA-9988)

Allow only one concurrent call to StatusLogger (CASSANDRA-12182)

Refactoring to specialised functional interfaces (CASSANDRA-13982)

Speculative retry should allow more friendly params (CASSANDRA-13876)

Throw exception if we send/receive repair messages to incompatible nodes (CASSANDRA-13944)

Replace usages of MessageDigest with Guava’s Hasher (CASSANDRA-13291)

Add nodetool cmd to print hinted handoff window (CASSANDRA-13728)

Fix some alerts raised by static analysis (CASSANDRA-13799)

Checksum sstable metadata (CASSANDRA-13321, CASSANDRA-13593)

Add result set metadata to prepared statement MD5 hash calculation (CASSANDRA-10786)

Add incremental repair support for --hosts, --force, and subrange repair (CASSANDRA-13818)

Refactor GcCompactionTest to avoid boxing (CASSANDRA-13941)

Expose recent histograms in JmxHistograms (CASSANDRA-13642)

Add SERIAL and LOCAL_SERIAL support for cassandra-stress (CASSANDRA-13925)

LCS needlessly checks for L0 STCS candidates multiple times (CASSANDRA-12961)

Update lz4 to 1.4.0 (CASSANDRA-13741)

Throttle base partitions during MV repair streaming to prevent OOM (CASSANDRA-13299)

Improve short read protection performance (CASSANDRA-13794)

Fix AssertionError in short read protection (CASSANDRA-13747)

Use compaction threshold for STCS in L0 (CASSANDRA-13861)

Fix problem with min_compress_ratio: 1 and disallow ratio < 1 (CASSANDRA-13703)

Add extra information to SASI timeout exception (CASSANDRA-13677)

Rework CompactionStrategyManager.getScanners synchronization (CASSANDRA-13786)

Add additional unit tests for batch behavior, TTLs, Timestamps (CASSANDRA-13846)

Add keyspace and table name in schema validation exception (CASSANDRA-13845)

Emit metrics whenever we hit tombstone failures and warn thresholds (CASSANDRA-13771)

Allow changing log levels via nodetool for related classes (CASSANDRA-12696)

Add stress profile yaml with LWT (CASSANDRA-7960)

Reduce memory copies and object creations when acting on ByteBufs (CASSANDRA-13789)

simplify mx4j configuration (Cassandra-13578)

Fix trigger example on 4.0 (CASSANDRA-13796)

force minimum timeout value (CASSANDRA-9375)

Add bytes repaired/unrepaired to nodetool tablestats (CASSANDRA-13774)

Don’t delete incremental repair sessions if they still have sstables (CASSANDRA-13758)

Fix pending repair manager index out of bounds check (CASSANDRA-13769)

Don’t use RangeFetchMapCalculator when RF=1 (CASSANDRA-13576)

Don’t optimise trivial ranges in RangeFetchMapCalculator (CASSANDRA-13664)

Use an ExecutorService for repair commands instead of new Thread(..).start() (CASSANDRA-13594)

Fix race / ref leak in anticompaction (CASSANDRA-13688)

Fix race / ref leak in PendingRepairManager (CASSANDRA-13751)

Enable ppc64le runtime as unsupported architecture (CASSANDRA-13615)

Improve sstablemetadata output (CASSANDRA-11483)

Support for migrating legacy users to roles has been dropped (CASSANDRA-13371)

Introduce error metrics for repair (CASSANDRA-13387)

Refactoring to primitive functional interfaces in AuthCache (CASSANDRA-13732)

Update metrics to 3.1.5 (CASSANDRA-13648)

batch_size_warn_threshold_in_kb can now be set at runtime (CASSANDRA-13699)

Avoid always rebuilding secondary indexes at startup (CASSANDRA-13725)

Upgrade JMH from 1.13 to 1.19 (CASSANDRA-13727)

Upgrade SLF4J from 1.7.7 to 1.7.25 (CASSANDRA-12996)

Default for start_native_transport now true if not set in config (CASSANDRA-13656)

Don’t add localhost to the graph when calculating where to stream from (CASSANDRA-13583)

Allow skipping equality-restricted clustering columns in ORDER BY clause (CASSANDRA-10271)

Use common nowInSec for validation compactions (CASSANDRA-13671)

Improve handling of IR prepare failures (CASSANDRA-13672)

Send IR coordinator messages synchronously (CASSANDRA-13673)

Flush system.repair table before IR finalize promise (CASSANDRA-13660)

Fix column filter creation for wildcard queries (CASSANDRA-13650)

Add 'nodetool getbatchlogreplaythrottle' and 'nodetool setbatchlogreplaythrottle' (CASSANDRA-13614)

fix race condition in PendingRepairManager (CASSANDRA-13659)

Allow noop incremental repair state transitions (CASSANDRA-13658)

Run repair with down replicas (CASSANDRA-10446)

Added started & completed repair metrics (CASSANDRA-13598)

Added started & completed repair metrics (CASSANDRA-13598)

Improve secondary index (re)build failure and concurrency handling (CASSANDRA-10130)

Improve calculation of available disk space for compaction (CASSANDRA-13068)

Change the accessibility of RowCacheSerializer for third party row cache plugins (CASSANDRA-13579)

Allow sub-range repairs for a preview of repaired data (CASSANDRA-13570)

NPE in IR cleanup when columnfamily has no sstables (CASSANDRA-13585)

Fix Randomness of stress values (CASSANDRA-12744)

Allow selecting Map values and Set elements (CASSANDRA-7396)

Fast and garbage-free Streaming Histogram (CASSANDRA-13444)

Update repairTime for keyspaces on completion (CASSANDRA-13539)

Add configurable upper bound for validation executor threads (CASSANDRA-13521)

Bring back maxHintTTL propery (CASSANDRA-12982)

Add testing guidelines (CASSANDRA-13497)

Add more repair metrics (CASSANDRA-13531)

RangeStreamer should be smarter when picking endpoints for streaming (CASSANDRA-4650)

Avoid rewrapping an exception thrown for cache load functions (CASSANDRA-13367)

Log time elapsed for each incremental repair phase (CASSANDRA-13498)

Add multiple table operation support to cassandra-stress (CASSANDRA-8780)

Cassandra is now relying on the JVM options to properly shutdown on OutOfMemoryError. By default it will rely on the OnOutOfMemoryError option as the ExitOnOutOfMemoryError and CrashOnOutOfMemoryError options are not supported by the older 1.7 and 1.8 JVMs. A warning will be logged at startup if none of those JVM options are used. See CASSANDRA-13006 for more details.

Creating Materialized View with filtering on non-primary-key base column (added in CASSANDRA-10368) is disabled, because the liveness of view row is depending on multiple filtered base non-key columns and base non-key column used in view primary-key. This semantic cannot be supported without storage format change, see CASSANDRA-13826. For append-only use case, you may still use this feature with a startup flag: "-Dcassandra.mv.allow_filtering_nonkey_columns_unsafe=true"

The NativeAccessMBean isAvailable method will only return true if the native library has been successfully linked. Previously it was returning true if JNA could be found but was not taking into account link failures.

Primary ranges in the system.size_estimates table are now based on the keyspace replication settings and adjacent ranges are no longer merged (CASSANDRA-9639).

In 2.1, the default for otc_coalescing_strategy was 'DISABLED'. In 2.2 and 3.0, it was changed to 'TIMEHORIZON', but that value was shown to be a performance regression. The default for 3.11.0 and newer has been reverted to 'DISABLED'. Users upgrading from Cassandra 2.2 or 3.0 should be aware that the default has changed.

The StorageHook interface has been modified to allow to retrieve read information from SSTableReader (CASSANDRA-13120).

Materialized Views for upgrades from DSE 5.1.1 or 5.1.2 or any version DSE 5.0.10 or later:

Runtime modification of concurrent_compactors is now available via nodetool concurrent_compactors.

Support for the assignment operators +=/-= has been added for update queries.

An Index implementation may now provide a task which runs prior to joining the ring. See CASSANDRA-12039

Filtering on partition key columns is now also supported for queries without secondary indexes.

A slow query log has been added: slow queries will be logged at DEBUG level. For more details refer to CASSANDRA-12403 and slow_query_log_timeout_in_ms in cassandra.yaml.

Support for GROUP BY queries has been added.

A new compaction-stress tool has been added to test the throughput of compaction for any cassandra-stress user schema. see compaction-stress help for how to use.

Prepared statements are now persisted in the table prepared_statements in the system keyspace. Upon startup, this table is used to preload all previously prepared statements - i.e. in many cases clients do not need to re-prepare statements against restarted nodes.

cqlsh can now connect to older Cassandra versions by downgrading the native protocol version. Please note that this is currently not part of our release testing and, as a consequence, it is not guaranteed to work in all cases. See CASSANDRA-12150 for more details.

Snapshots that are automatically taken before a table is dropped or truncated will have a "dropped" or "truncated" prefix on their snapshot tag name.

Metrics are exposed for successful and failed authentication attempts. These can be located using the object names org.apache.cassandra.metrics:type=Client,name=AuthSuccess and org.apache.cassandra.metrics:type=Client,name=AuthFailure respectively.

Add support to "unset" JSON fields in prepared statements by specifying DEFAULT UNSET. See CASSANDRA-11424 for details

Allow TTL with null value on insert and update. It will be treated as equivalent to inserting a 0.

Removed outboundBindAny configuration property. See CASSANDRA-12673 for details.

Support for alter types of already defined tables and of UDTs fields has been disabled. If it is necessary to return a different type, please use casting instead. See CASSANDRA-12443 for more details.

Specifying the default_time_to_live option when creating or altering a materialized view was erroneously accepted (and ignored). It is now properly rejected.

Only Java and JavaScript are now supported UDF languages. The sandbox in 3.0 already prevented the use of script languages except Java and JavaScript.

Compaction now correctly drops sstables out of CompactionTask when there isn’t enough disk space to perform the full compaction. This should reduce pending compaction tasks on systems with little remaining disk space.

Request timeouts in cassandra.yaml (read_request_timeout_in_ms, etc) now apply to the "full" request time on the coordinator. Previously, they only covered the time from when the coordinator sent a message to a replica until the time that the replica responded. Additionally, the previous behavior was to reset the timeout when performing a read repair, making a second read to fix a short read, and when subranges were read as part of a range scan or secondary index query. In 3.10 and higher, the timeout is no longer reset for these "subqueries". The entire request must complete within the specified timeout. As a consequence, your timeouts may need to be adjusted to account for this. See CASSANDRA-12256 for more details.

Logs written to stdout are now consistent with logs written to files. Time is now local (it was UTC on the console and local in files). Date, thread, file and line info where added to stdout. (see CASSANDRA-12004)

The 'clientutil' jar, which has been somewhat broken on the 3.x branch, is not longer provided. The features provided by that jar are provided by any good java driver and we advise relying on drivers rather on that jar, but if you need that jar for backward compatiblity until you do so, you should use the version provided on previous Cassandra branch, like the 3.0 branch (by design, the functionality provided by that jar are stable accross versions so using the 3.0 jar for a client connecting to 3.x should work without issues).

(Tools development) DatabaseDescriptor no longer implicitly startups components/services like commit log replay. This may break existing 3rd party tools and clients. In order to startup a standalone tool or client application, use the DatabaseDescriptor.toolInitialization() or DatabaseDescriptor.clientInitialization() methods. Tool initialization sets up partitioner, snitch, encryption context. Client initialization just applies the configuration but does not setup anything. Instead of using Config.setClientMode() or Config.isClientMode(), which are deprecated now, use one of the appropiate new methods in DatabaseDescriptor.

Application layer keep-alives were added to the streaming protocol to prevent idle incoming connections from timing out and failing the stream session (CASSANDRA-11839). This effectively deprecates the streaming_socket_timeout_in_ms property in favor of streaming_keep_alive_period_in_secs. See cassandra.yaml for more details about this property.

Duration literals support the ISO 8601 format. By consequence, identifiers matching that format (e.g P2Y or P1MT6H) will not be supported anymore (CASSANDRA-11873).

Shared pool threads are now named according to the stage they are executing tasks for. Thread names mentioned in traced queries change accordingly.

A new option has been added to cassandra-stress "-rate fixed={number}/s" that forces a scheduled rate of operations/sec over time. Using this, stress can accurately account for coordinated ommission from the stress process.

The cassandra-stress "-rate limit=" option has been renamed to "-rate throttle="

hdr histograms have been added to stress runs, it’s output can be saved to disk using: "-log hdrfile=" option. This histogram includes response/service/wait times when used with the fixed or throttle rate options. The histogram file can be plotted on http://hdrhistogram.github.io/HdrHistogram/plotFiles.html

TimeWindowCompactionStrategy has been added. This has proven to be a better approach to time series compaction and new tables should use this instead of DTCS. See CASSANDRA-9666 for details.

DateTieredCompactionStrategy has been deprecated - new tables should use TimeWindowCompactionStrategy. Note that migrating an existing DTCS-table to TWCS might cause increased compaction load for a while after the migration so make sure you run tests before migrating. Read CASSANDRA-9666 for background on this.

Change-Data-Capture is now available. See cassandra.yaml and for cdc-specific flags and a brief explanation of on-disk locations for archived data in CommitLog form. This can be enabled via ALTER TABLE …​ WITH cdc=true. Upon flush, CommitLogSegments containing data for CDC-enabled tables are moved to the data/cdc_raw directory until removed by the user and writes to CDC-enabled tables will be rejected with a WriteTimeoutException once cdc_total_space_in_mb is reached between unflushed CommitLogSegments and cdc_raw.

The ReversedType behaviour has been corrected for clustering columns of BYTES type containing empty value. Scrub should be run on the existing SSTables containing a descending clustering column of BYTES type to correct their ordering. See CASSANDRA-12127 for more details.

Ec2MultiRegionSnitch will no longer automatically set broadcast_rpc_address to the public instance IP if this property is defined on cassandra.yaml.

The name "json" and "distinct" are not valid anymore a user-defined function names (they are still valid as column name however). In the unlikely case where you had defined functions with such names, you will need to recreate those under a different name, change your code to use the new names and drop the old versions, and this before upgrade (see CASSANDRA-10783 for more details).

A maximum size for SSTables values has been introduced, to prevent out of memory exceptions when reading corrupt SSTables. This maximum size can be set via max_value_size_in_mb in cassandra.yaml. The default is 256MB, which matches the default value of native_transport_max_frame_size_in_mb. SSTables will be considered corrupt if they contain values whose size exceeds this limit. See CASSANDRA-9530 for more details.

JMX connections can now use the same auth mechanisms as CQL clients. New options in cassandra-env.(sh|ps1) enable JMX authentication and authorization to be delegated to the IAuthenticator and IAuthorizer configured in cassandra.yaml. The default settings still only expose JMX locally, and use the JVM’s own security mechanisms when remote connections are permitted. For more details on how to enable the new options, see the comments in cassandra-env.sh. A new class of IResource, JMXResource, is provided for the purposes of GRANT/REVOKE via CQL. See CASSANDRA-10091 for more details. Also, directly setting JMX remote port via the com.sun.management.jmxremote.port system property at startup is deprecated. See CASSANDRA-11725 for more details.

JSON timestamps are now in UTC and contain the timezone information, see CASSANDRA-11137 for more details.

Collision checks are performed when joining the token ring, regardless of whether the node should bootstrap. Additionally, replace_address can legitimately be used without bootstrapping to help with recovery of nodes with partially failed disks. See CASSANDRA-10134 for more details.

Key cache will only hold indexed entries up to the size configured by column_index_cache_size_in_kb in cassandra.yaml in memory. Larger indexed entries will never go into memory. See CASSANDRA-11206 for more details.

For tables having a default_time_to_live specifying a TTL of 0 will remove the TTL from the inserted or updated values.

Startup is now aborted if corrupted transaction log files are found. The details of the affected log files are now logged, allowing the operator to decide how to resolve the situation.

Filtering expressions are made more pluggable and can be added programatically via a QueryHandler implementation. See CASSANDRA-11295 for more details.

Internal authentication now supports caching of encrypted credentials. Reference cassandra.yaml:credentials_validity_in_ms

Remote configuration of auth caches via JMX can be disabled using the the system property cassandra.disable_auth_caches_remote_configuration

sstabledump tool is added to be 3.0 version of former sstable2json. The tool only supports v3.0+ SSTables. See tool’s help for more detail.

The mbean interfaces org.apache.cassandra.auth.PermissionsCacheMBean and org.apache.cassandra.auth.RolesCacheMBean are deprecated in favor of org.apache.cassandra.auth.AuthCacheMBean. This generalized interface is common across all caches in the auth subsystem. The specific mbean interfaces for each individual cache will be removed in a subsequent major version.

We now make sure that a token does not exist in several data directories. This means that we run one compaction strategy per data_file_directory and we use one thread per directory to flush. Use nodetool relocatesstables to make sure your tokens are in the correct place, or just wait and compaction will handle it. See CASSANDRA-6696 for more details.

bound maximum in-flight commit log replay mutation bytes to 64 megabytes tunable via cassandra.commitlog_max_outstanding_replay_bytes

Support for type casting has been added to the selection clause.

Hinted handoff now supports compression. Reference cassandra.yaml:hints_compression.

The Thrift API is deprecated and will be removed in Cassandra 4.0.

The compression ratio metrics computation has been modified to be more accurate.

Running Cassandra as root is prevented by default.

JVM options are moved from cassandra-env.(sh|ps1) to jvm.options.

The return value of SelectStatement::getLimit as been changed from DataLimits to int.

Custom index implementation should be aware that the method Indexer::indexes() has been removed as its contract was misleading and all custom implementation should have almost surely returned true inconditionally for that method.

GC logging is now enabled by default (you can disable it in the jvm.options file if you prefer).

Added a toString() serializer for GraphBinary.

Configured the Gremlin Console to use GraphBinary by default.

Fixed transaction management for empty iterators in Gremlin Server.

Deprecated MessageSerializer implementations for Gryo in Gremlin Server.

Deprecated Serializers enum values of GRYO_V1D0 and GRYO_V3D0.

Deprecated SerTokens values of MIME_GRYO_V1D0 and MIME_GRYO_V3D0.

Added a Docker command to start Gremlin Server with the standard GLV test configurations.

Added aggregate(Scope,String) and deprecated store() in favor of aggregate(local).

Modified NumberHelper to better ignore Double.NaN in min() and max() comparisons.

Bump to Netty 4.1.36.

Added userAgent to RequestOptions.

Gremlin Console sends Gremlin Console/version as the userAgent.

Fixed DriverRemoteConnection ignoring with Token options when multiple were set.

Added :set warnings true|false to Gremlin Console.

Provided support for withComputer() in gremlin-javascript.

Deprecated remote traversal side-effect retrieval and related infrastructure.

Bump to Jackson Databind 2.9.9.1.

Fixed bug with Python in g:Date of GraphSON where local time zone was being used during serialization/deserialization.

Deprecated multi/meta-property support in Neo4jGraph.

Improved exception and messaging for gt/gte/lt/lte when one of the object isn’t a Comparable.

Added test infrastructure to check for storage iterator leak.

Fixed multiple iterator leaks in query processor.

Fixed optional() so that the child traversal is treated as local.

Changed default keep-alive time for driver to 3 minutes.

Fixed bug where server-side keep-alive was not always disabled when its setting was zero.

Added support for hasNext() in Javascript and .NET.

Improved error messaging for invalid inputs to the TinkerGraph IdManager instances.

Forced replacement of connections in Java driver for certain exception types that seem to ultimately kill the connection.

Changed the reverse() of desc and asc on Order to not use the deprecated decr and incr.

Fixed bug in MatchStep where the correct was not properly determined.

Fixed bug where client/server exception mismatch when server throw StackOverflowError.

Added underscore suffixed steps and tokens in Gremlin-Python that conflict with global function names.

Prevent exception when closing a session that doesn’t exist.

Allow predicates and traversals to be used as options in BranchStep.

Ensure only a single final response is sent to the client with Gremlin Server.

Deprecated ResponseHandlerContext with related infrastructure and folded its functionality into Context in Gremlin Server.

Improved performance of aggregate() by avoiding excessive calls to hasNext() when the barrier is empty.