01 Dec 10:34

v3.5.2

New features

Pika supports Redis transactions. #2124 @ForestLH
Pika adds a cache layer for hot and cold data separation, improving read performance. #2133 @chejinge@Mixficsol
Codis-Proxy supports dynamic configuration parameter modification. #2103 @dingxiaoshuai123
Completing Go Test cases. #2063@dingxiaoshuai123
Adds cache to improve compilation speed on CI. #2088@baerwang

bugfix

Fixed coredump issue when using SETRANGE command in Pika. #2125 @chejinge
Fixed full replication issue caused by deleting Clearreplicationid from binlog. #2136 @Mixficsol
Modified lock granularity to improve binlog writing performance in Pika. #2129 @wangshao1
Added redis-copy traffic replication tool. #2044 @Mixficsol
Fixed potential data overflow issue in complex data types member variables. #2016 @u6th9d
Fixed incorrect return value issue in decr command. #2092 @dingxiaoshuai123
Fixed issue where SETRANGE and SETBIT commands did not retain the original key's expiration time. #2095 @u6th9d

Assets 4

28 Sep 02:45

New features

Slow log Adds statistics on queue waiting time #1997 @wangshao1
ReplicationID is used for primary/secondary replication #1951 @Mixficsol
WAL uses the disablewal command to support dynamic shutdown #2015 @Mixficsol
The number of threads flushed and the number of threads compaction dynamically adjust into one #2014 @Tianpingan
The RocksDB version was upgraded to v8.3.3 #1999 @dingxiaoshuai123
Added documentation for configuring Pika in macOS environment #2003 @klboke
Added the ability to periodically print the length of the work queue to quickly locate problems when the queue is blocked #1978 @Tianpingan
Added an indicator to detect the entire cluster using a pika_exporter #1953 @chenbt-hz
Realize automatic registration of Pika service on K8s environment, and automatically register at startup, so as to realize self-organization of cluster #1931 @machinly

bugfix

Reduces unnecessary log printing by the exporter, reducing CPU utilization #1945 @Mixficsol
The sentinel mechanism has been upgraded to carry out logical deletion of primary nodes that fall offline #1949 @Mixficsol
Adjust the rate_limit parameter to fix the situation that RPS is 0 during pressure measurement #2009 @chejinge
Fixed the logical determination of empty path when traversing data files in INFODATA command #1996 @Mixficsol
Improved the description of some missing parts of the document #1962 @baerwang
Use make-j to improve build speed #1933 @xiezheng-XD
Fixed an issue where large burrs appeared on the Codis line #2016 @chejinge
Fixed an issue where tools could not be compiled in macOS #2011 @A2ureStone

Assets 4

24 Aug 04:01

Major features

1. Removal of Rsync

In v3.5.0, we removed Rsync and replaced it with a self-developed full synchronization scheme, which has the features of breakpoint continuation, rate limit, file inspection and so on

2. Compatibility with More Redis Commands

More Redis commands are compatible and supported in v3.5.0.

3. RocksDB Version Upgrade and Tiered Compression

In v3.5.0, we upgraded RocksDB to version v8.1.1 and implemented tiered compression.

4. Support for BlobDB:

In v3.5.0, we added support for BlobDB KV separation.

5. Cluster Mode based on Codis

In v3.5.0, we introduced the Codis Cluster Mode and added support for the command to migrate Codis slots.

6.Observability

In v3.5.0, we introduced the pika_exporter tool to monitor the observability metrics of Pika. Subsequently, we added several new metrics to enhance Pika's observability.

7.Cloud-Native Deployment

In v3.5.0, we introduced an MVP version of pika-operator, which focuses on enabling the quick deployment of a single-instance Pika service on Kubernetes.

8.Cross-platform compilation

In v3.5.0, Pika supports cross-platform compilation and can be compiled and used on MacOS, CentOS, and Ubuntu platforms.

Features

Added automatic rate limiting for "compact" operation to reduce its impact on the latency of upper-layer data read and write operations. #1374 @wanghenshui
Introduced the "aof_to_pika" toolset for data conversion. #1340 @Axlgrep
Implemented the printing of the Pika logo. #1787 @Mixficsol
Optimized certain code portions using clang-tidy. #1701 #1730 @longfar-ncy
Refactored C++98-style code to C++11-style following C++ coding guidelines. #1684 @chejinge
Introduced support for the "snappy" library. #1216 @kernelai
Added CLA (Contributor License Agreement) file. #1260 @kernelai
Extended memory-related configuration options to support units in K, M, or G. #1307 @lqxhub
Added libunwind as a dependency library. #1316 @kernelai
Included a compilation script for Pika. #1356 @lqxhub
Added "rate-limiter-bandwidth" option to pika.conf for rate limiting. #1272 @wanghenshui
Enhanced the "info" command by adding "redis_version" information. #1403 @wanghenshui
Provided CMake support for aof_to_pika tool. #1436 @A2ureStone
Introduced clang-formatted shell scripts. #1448 @lqxhub
Added a configuration option to disable the compilation of command docs. #1642 @tedli
Provided CMake support for benchmark_client, binlog_sender, manifest_generator, rdb_to_pika, txt_to_pika, pika_to_txt, and pika_port. #1451 @A2ureStone

Bugfixes

Fixed coredump caused by incompatible CPU instruction sets. #1812 @chejinge
Fixed abnormal exit issue in the Monitor command. #1804 @Mixficsol
Fixed incomplete configuration overwrite. #1800 @yaoyinnan
Fixed CTest tests on MacOS. #1789 #1721 #1782 @tedli @Mixficsol
Fixed loading master-run-id from the configuration file during process startup, which caused the need for re-syncing historical data on slave nodes. #1778 @luky116
Fixed format error in overrides option during Pika single test. #1768 @A2ureStone
Fixed command to retrieve INFO xxx (INFO data, INFO stats) by index instead of using INFO ALL. #1762 @yaoyinnan
Improved instantaneous_metric to collect network metrics every 5 seconds. #1757 @yaoyinnan
Improved config get to return complete parameters. #1593 @luky116
Fixed possible empty slot issue. #1712 @Mixficsol
Fixed some bugs related to version selection in the makefile. #1714 @luky116
Fixed bug in executing HSET for the first time. #1710 @Mixficsol
Fixed the issue of binlog offset on slave nodes being greater than master nodes, causing TrySync failure. #1681 @luky116
Fixed a NOAUTH error test, where any command would give an error when AUTH is required. #1680 @hqh-cell
Fixed the issue of binlog files not being automatically cleaned up when exceeding the limit. #1679 @luky116
Fixed issues with LPUSHX and RPUSHX multi-element insertion. #1525 @ForestLH
Fixed error in executing the BGSAVE command. #1523 @Brokenice0415
Fixed the use of io_uring when closing RocksDB to prevent errors during linking. #1489 @lqxhub
Fixed MacOS environment Pika compilation warnings. #1740 @Mixficsol
Fixed error message for unsupported commands, adapted for go-redis. #1244 @wgqi1126
Fixed a protobuf3 compilation warning. #1267 @wanghenshui
Fixed bugs in EXISTS and DEL commands in sharding mode. #1277 @wanghenshui
Fixed rsync auth bug. #1278 @wanghenshui
Fixed rewrite result format to comply with Redis standards. #1339 @lqxhub
Fixed SET NX/XX return values to match Redis behavior. #1343 @gtygo
Fixed missing pika_master_link_status and pika_slave_priority indicators in exporter. #1726 @Mixficsol
Fixed potential coredump when Pika uses too many file descriptors, preventing it from exceeding the ulimit. #1346 @kernelai
Fixed initialization of group members in classes to prevent unexpected closure of the connection between redis-cli and Pika. #1390 @AlexStocks
Fixed delete_dir snprintf bug. #1400 @wanghenshui
Fixed thread num limitation. #1401 @wanghenshui
Fixed redis-benchmark to get config from Pika. #1402 @wanghenshui
Fixed initialization of class members. #1406 @AlexStocks
Fixed issues in Pika's master-slave synchronization tests on MacOS. #1776 @Mixficsol
Fixed NULL to nullptr. [#...

Assets 4

05 Aug 02:31

Major features

1. Removal of Rsync

In v3.5.0-alpha, we removed Rsync and replaced it with a self-developed full synchronization scheme, which has the features of breakpoint continuation, rate limit, file inspection and so on

2. Compatibility with More Redis Commands

More Redis commands are compatible and supported in v3.5.0-alpha.

3. RocksDB Version Upgrade and Tiered Compression

In v3.5.0-alpha, we upgraded RocksDB to version v8.1.1 and implemented tiered compression.

4. Support for BlobDB:

In v3.5.0-alpha, we added support for BlobDB KV separation.

5. Cluster Mode based on Codis

In v3.5.0-alpha, we introduced the Codis Cluster Mode and added support for the command to migrate Codis slots.

6.Observability

In v3.5.0-alpha, we introduced the pika_exporter tool to monitor the observability metrics of Pika. Subsequently, we added several new metrics to enhance Pika's observability.

7.Cloud-Native Deployment

In v3.5.0-alpha, we introduced an MVP version of pika-operator, which focuses on enabling the quick deployment of a single-instance Pika service on Kubernetes.

8.Cross-platform compilation

In v3.5.0-alpha, Pika supports cross-platform compilation and can be compiled and used on MacOS, CentOS, and Ubuntu platforms.

Features

Added automatic rate limiting for "compact" operation to reduce its impact on the latency of upper-layer data read and write operations. #1374 @wanghenshui
Introduced the "aof_to_pika" toolset for data conversion. #1340 @Axlgrep
Implemented the printing of the Pika logo. #1787 @Mixficsol
Optimized certain code portions using clang-tidy. #1701 #1730 @longfar-ncy
Refactored C++98-style code to C++11-style following C++ coding guidelines. #1684 @chejinge
Introduced support for the "snappy" library. #1216 @kernelai
Added CLA (Contributor License Agreement) file. #1260 @kernelai
Extended memory-related configuration options to support units in K, M, or G. #1307 @lqxhub
Added libunwind as a dependency library. #1316 @kernelai
Included a compilation script for Pika. #1356 @lqxhub
Added "rate-limiter-bandwidth" option to pika.conf for rate limiting. #1272 @wanghenshui
Enhanced the "info" command by adding "redis_version" information. #1403 @wanghenshui
Provided CMake support for aof_to_pika tool. #1436 @A2ureStone
Introduced clang-formatted shell scripts. #1448 @lqxhub
Added a configuration option to disable the compilation of command docs. #1642 @tedli
Provided CMake support for benchmark_client, binlog_sender, manifest_generator, rdb_to_pika, txt_to_pika, pika_to_txt, and pika_port. #1451 @A2ureStone

Bugfixes

Fixed coredump caused by incompatible CPU instruction sets. #1812 @chejinge
Fixed abnormal exit issue in the Monitor command. #1804 @Mixficsol
Fixed incomplete configuration overwrite. #1800 @yaoyinnan
Fixed CTest tests on MacOS. #1789 #1721 #1782 @tedli @Mixficsol
Fixed loading master-run-id from the configuration file during process startup, which caused the need for re-syncing historical data on slave nodes. #1778 @luky116
Fixed format error in overrides option during Pika single test. #1768 @A2ureStone
Fixed command to retrieve INFO xxx (INFO data, INFO stats) by index instead of using INFO ALL. #1762 @yaoyinnan
Improved instantaneous_metric to collect network metrics every 5 seconds. #1757 @yaoyinnan
Improved config get to return complete parameters. #1593 @luky116
Fixed possible empty slot issue. #1712 @Mixficsol
Fixed some bugs related to version selection in the makefile. #1714 @luky116
Fixed bug in executing HSET for the first time. #1710 @Mixficsol
Fixed the issue of binlog offset on slave nodes being greater than master nodes, causing TrySync failure. #1681 @luky116
Fixed a NOAUTH error test, where any command would give an error when AUTH is required. #1680 @hqh-cell
Fixed the issue of binlog files not being automatically cleaned up when exceeding the limit. #1679 @luky116
Fixed issues with LPUSHX and RPUSHX multi-element insertion. #1525 @ForestLH
Fixed error in executing the BGSAVE command. #1523 @Brokenice0415
Fixed the use of io_uring when closing RocksDB to prevent errors during linking. #1489 @lqxhub
Fixed MacOS environment Pika compilation warnings. #1740 @Mixficsol
Fixed error message for unsupported commands, adapted for go-redis. #1244 @wgqi1126
Fixed a protobuf3 compilation warning. #1267 @wanghenshui
Fixed bugs in EXISTS and DEL commands in sharding mode. #1277 @wanghenshui
Fixed rsync auth bug. #1278 @wanghenshui
Fixed rewrite result format to comply with Redis standards. #1339 @lqxhub
Fixed SET NX/XX return values to match Redis behavior. #1343 @gtygo
Fixed missing pika_master_link_status and pika_slave_priority indicators in exporter. #1726 @Mixficsol
Fixed potential coredump when Pika uses too many file descriptors, preventing it from exceeding the ulimit. #1346 @kernelai
Fixed initialization of group members in classes to prevent unexpected closure of the connection between redis-cli and Pika. #1390 @AlexStocks
Fixed delete_dir snprintf bug. #1400 @wanghenshui
Fixed thread num limitation. #1401 @wanghenshui
Fixed redis-benchmark to get config from Pika. #1402 @wanghenshui
Fixed initialization of class members. #1406 @AlexStocks
Fixed issues in Pika's master-slave synchronization tests on MacOS. #1776 [@Mixficsol](https://gi...

Assets 4

07 May 10:46

v3.4.1[abandoned]

This version has many bugs and is unstable, we have abandoned it. Use v3.3.6 instead.

新增功能：

支持quit命令。
支持gcc 9.4.0 编译。
多阶段编译减少docker image size。
release包对zlib、lz4、zstd压缩算法支持。

bugfix

修复线程池惊群问题。
修复pubsub 出现coredump的问题。
修复max-write-buffer-num 无法静态配置的问题。
修复max-cache-statistic-keys 配置项重复的问题。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。

Assets 4

01 Dec 03:51

v3.4.0[abandoned]

This version has many bugs and is unstable, we have abandoned it. Use v3.3.6 instead.

注意：

这是个实验性质的版本，360内部没有在生产环境部署。请优先使用3.3.6版本。

新增功能：

sharding模式下内置pika proxy组件，自动代理客户端请求到响应slot节点。
proxy支持根据业务压力配置后端连接数。
proxy 支持后端连接自动保活机制。
proxy 支持slot 共享后端连接。
proxy 支持客户端pipline功能。
proxy 支持slot 主从切换，slot 数据迁移功能。
proxy 支持hash tag功能。用户可以通过hash tag 把key存储到指定的slot。
pika 支持protocol buf 管理接口。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。
3.3.0 以后的版本包含了日志复制一致性的功能，建议使用该功能的用户使用版本3.3.6。
为了保证服务的稳定，建议不使用日志复制一致性的用户升级至3.2.9。
目前停止对3.0版本的维护

Assets 2

23 Oct 08:37

Do not use the sharding mode. We have abandoned this mode.

新增功能：

增加在分片模式下计算slot时对hash tag的支持。
增加在sharding模式下支持pkscanrange 、pkrscanrange、lpushrpop命令。客户端需要指定hash tag来保证所有操作在同一个slot上。
增加rocksdb memtable 配置预分配内存大小的接口。

Bug修复：

修复分片模式下，配置slot num非2的幂次数目时，计算slot hash值与codis不一致的问题。
修复哨兵使用pubsub命令可能导致coredump的问题。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。
3.3.0 以后的版本包含了日志复制一致性的功能，建议使用该功能的用户使用本版本。
为了保证服务的稳定，建议不使用日志复制一致性的用户升级至3.2.9。
目前停止对3.0版本的维护

Assets 3

07 Jul 08:35

性能优化：

针对zcout/zrangebyscore/zremrangebyscore/zrank命令性能优化。感谢@yurongliao的贡献

Bug修复：

修复pink中holy thread错误关闭文件描述符导致rocksdb中sst文件 bad file description。建议3.2以后的版本升级到本版本。感谢@su47flying的贡献。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。
3.3.0 以后的版本包含了日志复制一致性的功能，建议使用该功能的用户使用本版本。
为了保证服务的稳定，建议不使用日志复制一致性的用户升级至3.2.9。
3.0仍会继续维护，目前已经彻底停止对2.X的支持

Assets 3

12 Jun 02:20

Bug修复：

修复命令info replication中lag 计算溢出的问题。
修复因网络丢包造成binlog同步阻塞未及时重连的问题。
修复网络割接场景下，同步不能自动恢复的问题。
修复分片模式下，不支持大写pkcluster系列命令的问题。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。
3.3.0 以后的版本包含了日志复制一致性的功能，建议使用该功能的用户使用本版本。
为了保证服务的稳定，建议不使用日志复制一致性的用户升级至3.2.9。
3.0仍会继续维护，目前已经彻底停止对2.X的支持

Assets 3

05 Jun 10:15

Rocksdb更新：

在v3.2.9 基础上对Rocksdb 进行更新，Rocksdb 版本从5.9.2 升级至5.18.3，升级后的Rocksdb更加稳定。同时，所有原Pika接口保持不变，所有原Blackwidow 接口保持不变，继承3.2.9 以及之前所有优化。

注意事项：

分片模式和经典模式不可兼容，请在启动时候配置好启动模式
副本一致性可配置功能目前只支持分片模式。
分片模式下取消slaveof 命令，使用pkcluster slotsslaveof 替代，详细见Pika分片命令
经典模式下取消info log命令，info log 的binlog offset移至info replication，info log 的binlog size移至info data
分片模式下取消info replication命令，用pkcluster info slot替代
由于redis-cli 对于数据展示格式限制，对于pkcluster info slot的数据展示格式不够人性化。可以自行修改redis-cli代码 redis-cli modification
pika从2.1.4推荐需要使用gcc 4.8+版本编译，更新gcc后执行make distclean && make编译
pika3.1.0版本之后不再支持双主
pika3.1.0版本使用pb协议进行内部通信，不能直接和之前的版本建立主从关系，由低版本升级到pika3.1.0可以参照wiki进行升级(如何升级到Pika3.0、如何升级到Pika3.1)
由于zset精度的优化（自pika3.0.0起zset精度已与redis一致），如果你在低版本的pika（<3.0.0）中使用了geo功能，请在将其升级到pika3.0时不要直接使用nemo_to_blackwidow工具进行geo相关zset（其它结构的数据不受影响）数据的迁移，否则由于精度差异问题，迁移后的geo数据将损坏，建议使用客户端将geo相关zset数据重新导入。
为了保证服务的稳定，建议所有用户升级至3.2.9
3.0仍会继续维护，目前已经彻底停止对2.X的支持

Assets 3