Poll monitor socket with timeout=0 one last time after timeout expires.
This avoids heartbeat timeouts and connection churn on Lambda and other FaaS envs.
Stop sending $readPreference to standalone servers.
Stop sending $readPreference primary because it's the server default.
Remove outdated secondary_ok flag.
Resync command monitoring and unified test format tests.
Redact entire hello response when the command started contained speculativeAuthenticate.
Make OP_REPLY cursor.cursor_id always be an Int64.
Tweak spec test because pymongo unpins cursors eagerly after errors.
Tweak spec test for PoolClearedEvent ordering when MongoDB handshake fails (see DRIVERS-1785).
Only skip killCursors for some error codes.
Rely on SDAM error handling to close the connection after a state change error.
Add service_id to various events.
Retain reference to pinned sockets to prevent premptive closure by CPython's cyclic GC.
MongoClient now requires 2 connections and 2 threads to each MongoDB 4.4+ server.
With one connection, the server streams (or pushes) updated heartbeat info.
With the other connection, the client periodically pings the server to
establish an accurate round-trip time (RTT). This change optimizes the
discovery of server state changes such as replica set elections.
Additional changes:
- Mark server Unknown before retrying isMaster check.
- Always reset the pool _after_ marking the server unknown.
- Configure fail point before creating the client in test SpecRunner.
- Unfreeze with replSetFreeze:0 to ensure a speedy elections in test suite.
Transient errors inside transaction unpins the session.
Add brief docs about sharded transactions and add 3.9 changelog.
Tests changes:
Add spec tests for sharded transaction recoveryToken.
Speed up txn tests by reducing SDAM waiting time after a network error.
Remove outdated test workaround for killAllSessions.
Replace OP_INSERT/OP_UPDATE/OP_DELETE usage with OP_MSG with moreToCome
flag when the server is 3.6 or greater.
Unack'ed OP_MSG supports bypass_document_validation.
getLastError updatedExisting is always false on mongos after an OP_MSG
unack'ed update.
On Python 2, preallocate a bytearray and use slice assignment instead of
using str +=.
On Python 3, preallocate a bytearray and use a memoryview with recv_into.
Add retryWrites MongoClient and URI setting.
Retry once for supported single-statement write operations including:
insert_one, update_one, delete_one, find_one_and_update, find_one_and_replace,
find_one_and_delete, update and remove with multi=False, and find_and_modify.
Use _OpReply class instead of passing bytes around.
Remove unnecessary operation argument to receive message.
Move _first_batch to message.py to avoid circular import.
