Skip to content

Latest commit

 

History

History
689 lines (599 loc) · 24.8 KB

CHANGELOG.rst

File metadata and controls

689 lines (599 loc) · 24.8 KB

2.5.1

April 23, 2015

Bug Fixes

  • Fix thread safety in DC-aware load balancing policy (PYTHON-297)
  • Fix race condition in node/token rebuild (PYTHON-298)
  • Set and send serial consistency parameter (PYTHON-299)

2.5.0

March 30, 2015

Features

  • Integrated cqlengine object mapping package
  • Utility functions for converting timeuuids and datetime (PYTHON-99)
  • Schema metadata fetch window randomized, config options added (PYTHON-202)
  • Support for new Date and Time Cassandra types (PYTHON-190)

Bug Fixes

  • Fix index target for collection indexes (full(), keys()) (PYTHON-222)
  • Thread exception during GIL cleanup (PYTHON-229)
  • Workaround for rounding anomaly in datetime.utcfromtime (Python 3.4) (PYTHON-230)
  • Normalize text serialization for lookup in OrderedMap (PYTHON-231)
  • Support reading CompositeType data (PYTHON-234)
  • Preserve float precision in CQL encoding (PYTHON-243)

2.1.4

January 26, 2015

Features

  • SaslAuthenticator for Kerberos support (PYTHON-109)
  • Heartbeat for network device keepalive and detecting failures on idle connections (PYTHON-197)
  • Support nested, frozen collections for Cassandra 2.1.3+ (PYTHON-186)
  • Schema agreement wait bypass config, new call for synchronous schema refresh (PYTHON-205)
  • Add eventlet connection support (PYTHON-194)

Bug Fixes

  • Schema meta fix for complex thrift tables (PYTHON-191)
  • Support for 'unknown' replica placement strategies in schema meta (PYTHON-192)
  • Resolve stream ID leak on set_keyspace (PYTHON-195)
  • Remove implicit timestamp scaling on serialization of numeric timestamps (PYTHON-204)
  • Resolve stream id collision when using SASL auth (PYTHON-210)
  • Correct unhexlify usage for user defined type meta in Python3 (PYTHON-208)

2.1.3

December 16, 2014

Features

  • INFO-level log confirmation that a connection was opened to a node that was marked up (PYTHON-116)
  • Avoid connecting to peer with incomplete metadata (PYTHON-163)
  • Add SSL support to gevent reactor (PYTHON-174)
  • Use control connection timeout in wait for schema agreement (PYTHON-175)
  • Better consistency level representation in unavailable+timeout exceptions (PYTHON-180)
  • Update schema metadata processing to accommodate coming schema modernization (PYTHON-185)

Bug Fixes

  • Support large negative timestamps on Windows (PYTHON-119)
  • Fix schema agreement for clusters with peer rpc_addres 0.0.0.0 (PYTHON-166)
  • Retain table metadata following keyspace meta refresh (PYTHON-173)
  • Use a timeout when preparing a statement for all nodes (PYTHON-179)
  • Make TokenAware routing tolerant of statements with no keyspace (PYTHON-181)
  • Update add_collback to store/invoke multiple callbacks (PYTHON-182)
  • Correct routing key encoding for composite keys (PYTHON-184)
  • Include compression option in schema export string when disabled (PYTHON-187)

2.1.2

October 16, 2014

Features

  • Allow DCAwareRoundRobinPolicy to be constructed without a local_dc, defaulting instead to the DC of a contact_point (PYTHON-126)
  • Set routing key in BatchStatement.add() if none specified in batch (PYTHON-148)
  • Improved feedback on ValueError using named_tuple_factory with invalid column names (PYTHON-122)

Bug Fixes

  • Make execute_concurrent compatible with Python 2.6 (PYTHON-159)
  • Handle Unauthorized message on schema_triggers query (PYTHON-155)
  • Make execute_concurrent compatible with Python 2.6 (github-197)
  • Pure Python sorted set in support of UDTs nested in collections (PYTON-167)
  • Support CUSTOM index metadata and string export (PYTHON-165)

2.1.1

September 11, 2014

Features

  • Detect triggers and include them in CQL queries generated to recreate the schema (github-189)
  • Support IPv6 addresses (PYTHON-144) (note: basic functionality added; Windows platform not addressed (PYTHON-20))

Bug Fixes

  • Fix NetworkTopologyStrategy.export_for_schema (PYTHON-120)
  • Keep timeout for paged results (PYTHON-150)

Other

  • Add frozen<> type modifier to UDTs and tuples to handle CASSANDRA-7857

2.1.0

August 7, 2014

Bug Fixes

  • Correctly serialize and deserialize null values in tuples and user-defined types (PYTHON-110)
  • Include additional header and lib dirs, allowing libevwrapper to build against Homebrew and Mac Ports installs of libev (PYTHON-112 and 804dea3)

2.1.0c1

July 25, 2014

Bug Fixes

  • Properly specify UDTs for columns in CREATE TABLE statements
  • Avoid moving retries to a new host when using request ID zero (PYTHON-88)
  • Don't ignore fetch_size arguments to Statement constructors (github-151)
  • Allow disabling automatic paging on a per-statement basis when it's enabled by default for the session (PYTHON-93)
  • Raise ValueError when tuple query parameters for prepared statements have extra items (PYTHON-98)
  • Correctly encode nested tuples and UDTs for non-prepared statements (PYTHON-100)
  • Raise TypeError when a string is used for contact_points (github #164)
  • Include User Defined Types in KeyspaceMetadata.export_as_string() (PYTHON-96)

Other

  • Return list collection columns as python lists instead of tuples now that tuples are a specific Cassandra type

2.1.0b1

July 11, 2014

This release adds support for Cassandra 2.1 features, including version 3 of the native protocol.

Features

  • When using the v3 protocol, only one connection is opened per-host, and throughput is improved due to reduced pooling overhead and lock contention.
  • Support for user-defined types (Cassandra 2.1+)
  • Support for tuple type in (limited usage Cassandra 2.0.9, full usage in Cassandra 2.1)
  • Protocol-level client-side timestamps (see Session.use_client_timestamp)
  • Overridable type encoding for non-prepared statements (see Session.encoders)
  • Configurable serial consistency levels for batch statements
  • Use io.BytesIO for reduced CPU consumption (github #143)
  • Support Twisted as a reactor. Note that a Twisted-compatible API is not exposed (so no Deferreds), this is just a reactor implementation. (github #135, PYTHON-8)

Bug Fixes

  • Fix references to xrange that do not go through "six" in libevreactor and geventreactor (github #138)
  • Make BoundStatements inherit fetch_size from their parent PreparedStatement (PYTHON-80)
  • Clear reactor state in child process after forking to prevent errors with multiprocessing when the parent process has connected a Cluster before forking (github #141)
  • Don't share prepared statement lock across Cluster instances
  • Format CompositeType and DynamicCompositeType columns correctly in CREATE TABLE statements.
  • Fix cassandra.concurrent behavior when dealing with automatic paging (PYTHON-81)
  • Properly defunct connections after protocol errors
  • Avoid UnicodeDecodeError when query string is unicode (PYTHON-76)
  • Correctly capture dclocal_read_repair_chance for tables and use it when generating CREATE TABLE statements (PYTHON-84)
  • Avoid race condition with AsyncoreConnection that may cause messages to fail to be written until a new message is pushed
  • Make sure cluster.metadata.partitioner and cluster.metadata.token_map are populated when all nodes in the cluster are included in the contact points (PYTHON-90)
  • Make Murmur3 hash match Cassandra's hash for all values (PYTHON-89, github #147)
  • Don't attempt to reconnect to hosts that should be ignored (according to the load balancing policy) when a notification is received that the host is down.
  • Add CAS WriteType, avoiding KeyError on CAS write timeout (PYTHON-91)

2.0.2

June 10, 2014

Bug Fixes

  • Add six to requirements.txt
  • Avoid KeyError during schema refresh when a keyspace is dropped and TokenAwarePolicy is not in use
  • Avoid registering multiple atexit cleanup functions when the asyncore event loop is restarted multiple times
  • Delay initialization of reactors in order to avoid problems with shared state when using multiprocessing (PYTHON-60)
  • Add python-six to debian dependencies, move python-blist to recommends
  • Fix memory leak when libev connections are created and destroyed (github #93)
  • Ensure token map is rebuilt when hosts are removed from the cluster

2.0.1

May 28, 2014

Bug Fixes

  • Fix check for Cluster.is_shutdown in in @run_in_executor decorator

2.0.0

May 28, 2014

Features

  • Make libev C extension Python3-compatible (PYTHON-70)
  • Support v2 protocol authentication (PYTHON-73, github #125)

Bug Fixes

  • Fix murmur3 C extension compilation under Python3.4 (github #124)

Merged From 1.x

Features

  • Add Session.default_consistency_level (PYTHON-14)

Bug Fixes

  • Don't strip trailing underscores from column names when using the named_tuple_factory (PYTHON-56)
  • Ensure replication factors are ints for NetworkTopologyStrategy to avoid TypeErrors (github #120)
  • Pass WriteType instance to RetryPolicy.on_write_timeout() instead of the string name of the write type. This caused write timeout errors to always be rethrown instead of retrying. (github #123)
  • Avoid submitting tasks to the ThreadPoolExecutor after shutdown. With retries enabled, this could cause Cluster.shutdown() to hang under some circumstances.
  • Fix unintended rebuild of token replica map when keyspaces are discovered (on startup), added, or updated and TokenAwarePolicy is not in use.
  • Avoid rebuilding token metadata when cluster topology has not actually changed
  • Avoid preparing queries for hosts that should be ignored (such as remote hosts when using the DCAwareRoundRobinPolicy) (PYTHON-75)

Other

  • Add 1 second timeout to join() call on event loop thread during interpreter shutdown. This can help to prevent the process from hanging during shutdown.

2.0.0b1

May 6, 2014

Upgrading from 1.x

Cluster.shutdown() should always be called when you are done with a Cluster instance. If it is not called, there are no guarantees that the driver will not hang. However, if you do have a reproduceable case where Cluster.shutdown() is not called and the driver hangs, please report it so that we can attempt to fix it.

If you're using the 2.0 driver against Cassandra 1.2, you will need to set your protocol version to 1. For example:

cluster = Cluster(..., protocol_version=1)

Features

  • Support v2 of Cassandra's native protocol, which includes the following new features: automatic query paging support, protocol-level batch statements, and lightweight transactions
  • Support for Python 3.3 and 3.4
  • Allow a default query timeout to be set per-Session

Bug Fixes

  • Avoid errors during interpreter shutdown (the driver attempts to cleanup daemonized worker threads before interpreter shutdown)

Deprecations

The following functions have moved from cassandra.decoder to cassandra.query. The original functions have been left in place with a DeprecationWarning for now:

  • cassandra.decoder.tuple_factory has moved to cassandra.query.tuple_factory
  • cassandra.decoder.named_tuple_factory has moved to cassandra.query.named_tuple_factory
  • cassandra.decoder.dict_factory has moved to cassandra.query.dict_factory
  • cassandra.decoder.ordered_dict_factory has moved to cassandra.query.ordered_dict_factory

Exceptions that were in cassandra.decoder have been moved to cassandra.protocol. If you handle any of these exceptions, you must adjust the code accordingly.

1.1.2

May 8, 2014

Features

  • Allow a specific compression type to be requested for communications with Cassandra and prefer lz4 if available

Bug Fixes

  • Update token metadata (for TokenAware calculations) when a node is removed from the ring
  • Fix file handle leak with gevent reactor due to blocking Greenlet kills when closing excess connections
  • Avoid handling a node coming up multiple times due to a reconnection attempt succeeding close to the same time that an UP notification is pushed
  • Fix duplicate node-up handling, which could result in multiple reconnectors being started as well as the executor threads becoming deadlocked, preventing future node up or node down handling from being executed.
  • Handle exhausted ReconnectionPolicy schedule correctly

Other

  • Don't log at ERROR when a connection is closed during the startup communications
  • Mke scales, blist optional dependencies

1.1.1

April 16, 2014

Bug Fixes

  • Fix unconditional import of nose in setup.py (github #111)

1.1.0

April 16, 2014

Features

  • Gevent is now supported through monkey-patching the stdlib (PYTHON-7, github issue #46)
  • Support static columns in schemas, which are available starting in Cassandra 2.1. (github issue #91)
  • Add debian packaging (github issue #101)
  • Add utility methods for easy concurrent execution of statements. See the new cassandra.concurrent module. (github issue #7)

Bug Fixes

  • Correctly supply compaction and compression parameters in CREATE statements for tables when working with Cassandra 2.0+
  • Lowercase boolean literals when generating schemas
  • Ignore SSL_ERROR_WANT_READ and SSL_ERROR_WANT_WRITE socket errors. Previously, these resulted in the connection being defuncted, but they can safely be ignored by the driver.
  • Don't reconnect the control connection every time Cluster.connect() is called
  • Avoid race condition that could leave ResponseFuture callbacks uncalled if the callback was added outside of the event loop thread (github issue #95)
  • Properly escape keyspace name in Session.set_keyspace(). Previously, the keyspace name was quoted, but any quotes in the string were not escaped.
  • Avoid adding hosts to the load balancing policy before their datacenter and rack information has been set, if possible.
  • Avoid KeyError when updating metadata after droping a table (github issues #97, #98)
  • Use tuples instead of sets for DCAwareLoadBalancingPolicy to ensure equal distribution of requests

Other

  • Don't ignore column names when parsing typestrings. This is needed for user-defined type support. (github issue #90)
  • Better error message when libevwrapper is not found
  • Only try to import scales when metrics are enabled (github issue #92)
  • Cut down on the number of queries executing when a new Cluster connects and when the control connection has to reconnect (github issue #104, PYTHON-59)
  • Issue warning log when schema versions do not match

1.0.2

March 4, 2014

Bug Fixes

  • With asyncorereactor, correctly handle EAGAIN/EWOULDBLOCK when the message from Cassandra is a multiple of the read buffer size. Previously, if no more data became available to read on the socket, the message would never be processed, resulting in an OperationTimedOut error.
  • Double quote keyspace, table and column names that require them (those using uppercase characters or keywords) when generating CREATE statements through KeyspaceMetadata and TableMetadata.
  • Decode TimestampType as DateType. (Cassandra replaced DateType with TimestampType to fix sorting of pre-unix epoch dates in CASSANDRA-5723.)
  • Handle latest table options when parsing the schema and generating CREATE statements.
  • Avoid 'Set changed size during iteration' during query plan generation when hosts go up or down

Other

  • Remove ignored tracing_enabled parameter for SimpleStatement. The correct way to trace a query is by setting the trace argument to True in Session.execute() and Session.execute_async().
  • Raise TypeError instead of cassandra.query.InvalidParameterTypeError when a parameter for a prepared statement has the wrong type; remove cassandra.query.InvalidParameterTypeError.
  • More consistent type checking for query parameters
  • Add option to a return special object for empty string values for non-string columns

1.0.1

Feb 19, 2014

Bug Fixes

  • Include table indexes in KeyspaceMetadata.export_as_string()
  • Fix broken token awareness on ByteOrderedPartitioner
  • Always close socket when defuncting error'ed connections to avoid a potential file descriptor leak
  • Handle "custom" types (such as the replaced DateType) correctly
  • With libevreactor, correctly handle EAGAIN/EWOULDBLOCK when the message from Cassandra is a multiple of the read buffer size. Previously, if no more data became available to read on the socket, the message would never be processed, resulting in an OperationTimedOut error.
  • Don't break tracing when a Session's row_factory is not the default namedtuple_factory.
  • Handle data that is already utf8-encoded for UTF8Type values
  • Fix token-aware routing for tokens that fall before the first node token in the ring and tokens that exactly match a node's token
  • Tolerate null source_elapsed values for Trace events. These may not be set when events complete after the main operation has already completed.

Other

  • Skip sending OPTIONS message on connection creation if compression is disabled or not available and a CQL version has not been explicitly set
  • Add details about errors and the last queried host to OperationTimedOut

1.0.0 Final

Jan 29, 2014

Bug Fixes

  • Prevent leak of Scheduler thread (even with proper shutdown)
  • Correctly handle ignored hosts, which are common with the DCAwareRoundRobinPolicy
  • Hold strong reference to prepared statement while executing it to avoid garbage collection
  • Add NullHandler logging handler to the cassandra package to avoid warnings about there being no configured logger
  • Fix bad handling of nodes that have been removed from the cluster
  • Properly escape string types within cql collections
  • Handle setting the same keyspace twice in a row
  • Avoid race condition during schema agreement checks that could result in schema update queries returning before all nodes had seen the change
  • Preserve millisecond-level precision in datetimes when performing inserts with simple (non-prepared) statements
  • Properly defunct connections when libev reports an error by setting errno instead of simply logging the error
  • Fix endless hanging of some requests when using the libev reactor
  • Always start a reconnection process when we fail to connect to a newly bootstrapped node
  • Generators map to CQL lists, not key sequences
  • Always defunct connections when an internal operation fails
  • Correctly break from handle_write() if nothing was sent (asyncore reactor only)
  • Avoid potential double-erroring of callbacks when a connection becomes defunct

Features

  • Add default query timeout to Session
  • Add timeout parameter to Session.execute()
  • Add WhiteListRoundRobinPolicy as a load balancing policy option
  • Support for consistency level LOCAL_ONE
  • Make the backoff for fetching traces exponentially increasing and configurable

Other

  • Raise Exception if TokenAwarePolicy is used against a cluster using the Murmur3Partitioner if the murmur3 C extension has not been compiled
  • Add encoder mapping for OrderedDict
  • Use timeouts on all control connection queries
  • Benchmark improvements, including command line options and eay multithreading support
  • Reduced lock contention when using the asyncore reactor
  • Warn when non-datetimes are used for 'timestamp' column values in prepared statements
  • Add requirements.txt and test-requirements.txt
  • TravisCI integration for running unit tests against Python 2.6, Python 2.7, and PyPy

1.0.0b7

Nov 12, 2013

This release makes many stability improvements, especially around prepared statements and node failure handling. In particular, several cases where a request would never be completed (and as a result, leave the application hanging) have been resolved.

Features

  • Add timeout kwarg to ResponseFuture.result()
  • Create connection pools to all hosts in parallel when initializing new Sesssions.

Bug Fixes

  • Properly set exception on ResponseFuture when a query fails against all hosts
  • Improved cleanup and reconnection efforts when reconnection fails on a node that has recently come up
  • Use correct consistency level when retrying failed operations against a different host. (An invalid consistency level was being used, causing the retry to fail.)
  • Better error messages for failed Session.prepare() opertaions
  • Prepare new statements against all hosts in parallel (formerly sequential)
  • Fix failure to save the new current keyspace on connections. (This could cause problems for prepared statements and lead to extra operations to continuously re-set the keyspace.)
  • Avoid sharing LoadBalancingPolicies across Cluster instances. (When a second Cluster was connected, it effectively mark nodes down for the first Cluster.)
  • Better handling of failures during the re-preparation sequence for unrecognized prepared statements
  • Throttle trashing of underutilized connections to avoid trashing newly created connections
  • Fix race condition which could result in trashed connections being closed before the last operations had completed
  • Avoid preparing statements on the event loop thread (which could lead to deadlock)
  • Correctly mark up non-contact point nodes discovered by the control connection. (This lead to prepared statements not being prepared against those hosts, generating extra traffic later when the statements were executed and unrecognized.)
  • Correctly handle large messages through libev
  • Add timeout to schema agreement check queries
  • More complete (and less contended) locking around manipulation of the pending message deque for libev connections

Other

  • Prepare statements in batches of 10. (When many prepared statements are in use, this allows the driver to start utilizing nodes that were restarted more quickly.)
  • Better debug logging around connection management
  • Don't retain unreferenced prepared statements in the local cache. (If many different prepared statements were created, this would increase memory usage and greatly increase the amount of time required to begin utilizing a node that was added or marked up.)

1.0.0b6

Oct 22, 2013

Bug Fixes

  • Use lazy string formatting when logging
  • Avoid several deadlock scenarios, especially when nodes go down
  • Avoid trashing newly created connections due to insufficient traffic
  • Gracefully handle un-handled Exceptions when erroring callbacks

Other

  • Node state listeners (which are called when a node is added, removed, goes down, or comes up) should now be registered through Cluster.register_listener() instead of through a host's HealthMonitor (which has been removed)

1.0.0b5

Oct 10, 2013

Features

  • SSL support

Bug Fixes

  • Avoid KeyError when building replica map for NetworkTopologyStrategy
  • Work around python bug which causes deadlock when a thread imports the utf8 module
  • Handle no blist library, which is not compatible with pypy
  • Avoid deadlock triggered by a keyspace being set on a connection (which may happen automatically for new connections)

Other

  • Switch packaging from Distribute to setuptools, improved C extension support
  • Use PEP 386 compliant beta and post-release versions

1.0.0-beta4

Sep 24, 2013

Features

  • Handle new blob syntax in Cassandra 2.0 by accepting bytearray objects for blob values
  • Add cql_version kwarg to Cluster.__init__

Bug Fixes

  • Fix KeyError when building token map with NetworkTopologyStrategy keyspaces (this prevented a Cluster from successfully connecting at all).
  • Don't lose default consitency level from parent PreparedStatement when creating BoundStatements

1.0.0-beta3

Sep 20, 2013

Features

  • Support for LZ4 compression (Cassandra 2.0+)
  • Token-aware routing will now utilize all replicas for a query instead of just the first replica

Bug Fixes

  • Fix libev include path for CentOS
  • Fix varint packing of the value 0
  • Correctly pack unicode values
  • Don't attempt to return failed connections to the pool when a final result is set
  • Fix bad iteration of connection credentials
  • Use blist's orderedset for set collections and OrderedDict for map collections so that Cassandra's ordering is preserved
  • Fix connection failure on Windows due to unavailability of inet_pton and inet_ntop. (Note that IPv6 inet_address values are still not supported on Windows.)
  • Boolean constants shouldn't be surrounded by single quotes
  • Avoid a potential loss of precision on float constants due to string formatting
  • Actually utilize non-standard ports set on Cluster objects
  • Fix export of schema as a set of CQL queries

Other

  • Use cStringIO for connection buffer for better performance
  • Add __repr__ method for Statement classes
  • Raise InvalidTypeParameterError when parameters of the wrong type are used with statements
  • Make all tests compatible with Python 2.6
  • Add 1s timeout for opening new connections

1.0.0-beta2

Aug 19, 2013

Bug Fixes

  • Fix pip packaging

1.0.0-beta

Aug 16, 2013

Initial release