Skip to content
Fetching contributors…
Cannot retrieve contributors at this time
254 lines (204 sloc) 10.9 KB
Redis 2.4 release notes
Migrating from 2.2 to 2.4
Redis 2.2 is mostly a strict subset of 2.4.
The only thing you should be aware is that you can't use .rdb and AOF files
generated with 2.4 into a 2.2 instance.
2.4 slaves can be attached to 2.2 masters, but not the contrary, and only for
the time needed to perform the version upgrade.
From the point of view of the API Redis 2.4 only adds new commands
(other commands now accepts a variable number of arguments) so you don't need
to modify your program in order to use Redis 2.4.
What's new in Redis 2.4.6
* [BUGFIX] Fixed issue #141 part 1: Possible protocol desyncs when clients send
wrong protocol is now fixed. (See issue 141 for more details)
* [BUGFIX] Fixed issue #141 part 2: Connection of multiple slaves used to result
from time to time into corrupted protocol send to slaves connected
after the first one. (See issue 141 for more details)
* [BUGFIX] Do not propagate DEBUG LOADAOF.
* New INFO contains information such as ip/port/state for every conneced slave.
* Show GCC version in INFO output.
What's new in Redis 2.4.5
* [BUGFIX] Fixed a ZUNIONSTORE/ZINTERSTORE bug that can cause a NaN to be
inserted as a sorted set element score. This happens when one of the
elements has +inf/-inf score and the weight used is 0.
* [BUGFIX] Fixed memory leak in CLIENT INFO.
* [BUGFIX] Fixed a non critical SORT bug (Issue 224).
* [BUGFIX] Fixed a replication bug: now the timeout configuration is respected
during the connection with the master.
* --quiet option implemented in the Redis test.
What's new in Redis 2.4.4
* [BUGFIX] jemalloc upgraded to version 2.2.5, previous versions had a
potentially serious issue when allocating big memory areas, something that
Redis actually does. However we never received bug reports that appear
to be caused by jemalloc.
* [BUGFIX] DISCARD now clears DIRTY_CAS flag in the client. Now the next
transaction will not fail if the previous transaction used WATCH and
the key was touched.
* CLIENT LIST output modified to include the last command executed by clients.
* Better bug report on crash.
* Protocol errors are now logged for loglevel >= verbose.
* Two new INFO fields related to AOF, that can be useful when investigating
Redis issues.
What's new in Redis 2.4.3
* redis-cli now supports 'single quotes' style strings.
* It is possible to disable password auth with CONFIG SET.
* Some section of redis.conf better documented.
* Default timeout for client connections now is 0 (no timeout).
* I/O buffer length modified for better performances with big payloads.
* Fixed crash on SPARC due to improper alighment due to bad assumptions about data types size.
* CLIENT LIST output improved, code refactored.
* [BUGFIX] Max log message length set to 4k to avoid truncation in INFO output written in the log file after a crash.
* [BUGFIX] Close client connection when the query buffer reaches 1GB and log informaiton about the offending client, instead of crashing the instance when the query buffer reaches 2GB (for sds.c string overflow). This is related to issue #141 (github issues) and should fix the effect of this rare bug related to replication, but the cause is still not perfectly clear.
What's new in Redis 2.4.2
* [BUGFIX] Unix socket creation mask config directive fixed.
* [BUGFIX] Fixed a bug that forced Redis to continously rewrite the AOF file
when a manual BGREWRITEAOF was issued during a BGSAVE was
still in progress.
* [BUGFIX] Fixed a server crash resulting from bad handling of SLAVEOF NO ONE.
* [BUGFIX] Fixed a memory leak in redis-cli.
* Debian/Ubuntu Redis system-wide installation script added under /utils.
What's new in Redis 2.4.1
* [BUGFIX] FLUSHALL was not replicated nor written into the Append Only File.
* [BUGFIX] FLUSHALL now only performs a sync SAVE if there is at least
one save point configured.
What's new in Redis 2.4.0
* [BUGFIX] redis-cli segfault with single numerical argument fixed.
* [BUGFIX] OpenBSD compilation problem fixed.
* [BUGFIX] More robust Redis test, with better random port selection.
* [BUGFIX] Fix for bug #128 about the RENAME command.
* [BUGFIX] Fixed Issue #131. stime/utime reported in INFO was inverted.
* [BUGFIX] Unlink Unix socket file on shutdown.
* [BUGFIX] AUTH now returns error if no password is set on the server.
* [BUGFIX] Exit with Fatal error at startup on RDB loading errors.
* redis-check-dump: RDB version 2 now supported.
* More informative error when DEBUG RELOAD fails.
* Added a config directive for a Unix socket mask.
* CONFIG SET/GET for loglevel.
What's new in Redis 2.3.11 (2.4 Release Candidate 8)
* [BUGFIX] Fixed a rare but possible AOF race condition that could result into
duplicated commands inside the AOF.
* [BUGFIX] Fixed issue 620, don't segfault on corrupted (by hand) AOF.
* [BUGFIX] Fixed compilation on Mac/PPC.
* [BUGFIX] Don't replicate SAVE.
* LRANGE optimization may drastically improve performances when querying the
final part of a long list.
* redis-cli now implements a --latency mode to monitory Redis delay.
* Hash type settings removed from INFO (same info is available via config GET)
* Include port number on error when can't bind.
* AOF fsync is now performed in background when fsync policy is 'everysec'.
* AOF performances improved moving in background a possibly slow close(2) call.
* AOF protocol synthesis speedup.
What's new in Redis 2.3.10 (2.4 Release Candidate 7)
* [BUGFIX] Fixed issue 593 (BRPOPLPUSH related crash).
* [BUGFIX] Fixed an issue with the networking layer that may prevent Redis from sending the whole reply back to client under extreme conditions.
What's new in Redis 2.3.9 (2.4 Release Candidate 6)
* [BUGFIX] Fixed a bug with the automatic AOF rewrite causing continuous
rewrites for AOF files bigger than 4 GB.
* New maxmemory tests.
What's new in Redis 2.3.8 (2.4 Release Candidate 5)
* The Slowlog feature was backported to Redis 2.2.
* A number of fixes related blocking operations on lists when mixed with
AOF and Replication.
* Fixed bad interactions between EXPIRE, EXPIREAT, and in general volatile
keys when AOF is enabled. More details in the Redis Google Group here:
* no more allocation stats info in INFO.
* colorized make for 2.2 as well.
* Fixed a problem with AOF when it is stopped via CONFIG SET appendonly no.
* Warn the user enabling VM that VM is deprecated and discouraged.
* prepareForShutdown() fixed for correctness.
* Close the listening sockets on exit for faster restarts.
What's new in Redis 2.2.11
* Solved a never reported but possibly critical bug in the AOF and RDB
persistence, introduced with the new version of the iterator: In very rare
circumstances the AOF (after rerwite) or the rdb file may contain the same
key more than one time.
What's new in Redis 2.2.10
* Fix for issue 566. It was possible for SETNX to behave in abnormal ways.
* Work with make v3.80
* redis-cli updated to the same version as the 2.4 branch. Now the
reconnection handling is better. You can prefix commands with a number
to execute the command the specified number of times. When using -r
in non interactive mode it is possible to set a delay between commands.
* Better init script.
What's new in Redis 2.2.9
Redis 2.2.9 is an important bugfix release:
* A Slave used to rewrite the AOF log after a sync with the master, but the
rewrite was performed at the wrong time, causing inconsistencies in the
AOF file generated.
* [LR]LPUSH replication in presence of BRPOPLPUSH was broken. Fixed.
* Memory leak in BRPOPLPUSH fixed.
* Pub/Sub bug resulting in random crashes fixed.
What's new in Redis 2.2.8
* A new form of dict.c (hash table implementation) iterator that performs less
copy-on-write of pages, introduced in Redis 2.2.7, caused ZINTERSTORE,
ZUNIONSTORE, SINTER, SINTERSTORE commands to behave in the wrong way.
This bug is now fixed.
* Print version info before running the test with 'make test'. This is mainly
useful for the Continuous Integration system we run.
* Fix for DEBUG DIGEST, key may expire on lookup, producing the wrong result.
* Replication with expire test modified to produce no or less false failures.
* Fixed Z*STORE when dealing with intsets, regression test added.
What's new in Redis 2.2.7
* Fixed bug #543-2 (the issue was reopened with a completely different report)
that caused Redis to randomly crash on list push performed against lists
with other clients blocked with BLPOP (or variants).
What's new in Redis 2.2.6
* Fixed bug #543. If you saw Redis instances crashing on List operations
(only happening with a non-default max entry size ziplist setting in
redis.conf) it was almost certainly this problem.
* Fixed a bug with replication where SLAVEOF NO ONE caused a slave to close the
connection with all its slaves.
What's new in Redis 2.2.5
* Fixed a crash occurring when loading an AOF containing the SPOP command.
Compared to Redis 2.3.7 (RC4) there are the following changes:
* [BUGFIX] HDEL: Abort deleting fields when hash is removed.
* [BUGFIX] Fix adding bulk reply when getcwd fails.
WHAT'S NEW IN REDIS 2.4 compared to the 2.2 version?
* Specially encoded sorted sets, now small sorted sets will use little memory.
* Native persistence of specially encoded data types (ziplists, zipmaps,
intsets). Many data sets will be saved and loaded an order of magnitude
* Variadic versions of commands: SADD, HDEL, SREM, ZREM, ZADD, L/RPUSH.
* Jemalloc support (enabled by default for Linux build) in order to
avoid fragmentation issues.
* Reduced memory usage while saving.
* More info fields (peak memory, fork time, ...)
* OBJECT command for objects introspection.
* CLIENT command for clients introspection.
* Non blocking slave -> master connection.
* Better redis-cli connection handling. New redis-cli features.
* Better redis-benchmark, now able to benchmark user provided commands.
* Colorized Make.
* VM deprecated. Still supported but with a big warning... don't use it.
* Many speed optimizations and bug fixes.
Credits: Where not specified the implementation and design are done by
Salvatore Sanfilippo and Pieter Noordhuis. Thanks to VMware for making all
this possible. Also many thanks to all the other contributors and the amazing
community we have.
Jump to Line
Something went wrong with that request. Please try again.