Redis Enterprise 软件发行说明 6.2.10(2022 年 2 月)
Python 3 支持。RHEL 8.5 支持。
Redis 企业软件 |
---|
Redis 企业软件版本 6.2.10 现已推出!
下表显示了可用软件包的 MD5 校验和。
包 | MD5 校验和 |
---|---|
Ubuntu 16 版本 | 531cea69a58fbc1125bc5f76ba01da7f |
Ubuntu 18 的 | ec9ac6e0111dc85605d3b98e83f50150 |
RedHat Enterprise Linux (RHEL) 7 Oracle Enterprise Linux (OL) 7 |
2f7572caab9600417ef8b4ee474d6768 |
RedHat Enterprise Linux (RHEL) 8 Oracle Enterprise Linux (OL) 8 |
377a539ee050515e1e0640dec1e04129 |
K8s Ubuntu | 099192416a70a12790535bdcd78a6e87 |
K8s RHEL | f267abe81770ddf36f022232f4c2cb2e |
功能和增强功能
-
将 Redis Enterprise 基础架构升级到 Python v3.9。
-
Red Hat Enterprise Linux (RHEL) v8.5 和 Red Hat Enterprise Linux (RHEL) v8.6 现在是受支持的平台。
-
与开源 Redis 6.2.5 兼容。
-
与最新开源 Redis 6.2.6 的安全修复程序兼容。
-
增强功能和错误修复。
版本变更
先决条件和说明
-
您可以从 Redis Enterprise Software v6.0 及更高版本升级到 v6.2.10。
-
请参阅 v6.2.4 发行说明,了解有关升级所做更改的重要说明。
-
不支持从 v6.0 之前的版本升级。
-
如果您计划将集群升级到 RHEL 8,请参阅 v6.2.8 发行说明了解已知限制。
-
如果您使用的是主动-主动或主动-被动 (ReplicaOf) 数据库,并且由于升级而遇到同步问题,请参阅已解决的问题中的RS67434详细信息,以帮助解决问题。
产品生命周期更新
Redis Enterprise Software v6.0.x 将于 2022 年 5 月 31 日终止生命周期 (EOF)。
要了解更多信息,请参阅 Redis Enterprise Software 产品生命周期,其中详细介绍了 Redis Enterprise Software 的版本号和生命周期终止计划。
有关 Redis 模块的信息和生命周期,请参阅模块生命周期。
Redis 模块
Redis Enterprise Software v6.2.10 包括以下 Redis 模块:
- 雷迪搜索 v2.2.6
- RedisJSON 版本 v2.0.6
- RedisBloom 版本 v2.2.9
- RedisGraph 版本 v2.4.12
- RedisTimeSeries 版本 v1.4.13
从 Redis Enterprise Software v6.2.10 build 121 开始,包含的模块版本为:
- RediSearch v2.4.6 版本
- RedisJSON 版本 v2.0.8
- RedisBloom 版本 2.2.14
- RedisGraph 版本 v2.8.12
- RedisTimeSeries 版本 v1.6.9
有关升级模块的帮助,请参阅将模块添加到集群。
界面增强功能
- 选择 RedisJSON 时,用户界面 (UI) 现在也会建议 RedisSearch。要了解更多信息,请参阅 RedisJSON 预览公告,其中详细介绍了结合使用 RedisJSON 和 RediSearch 的好处。
- 添加对节点列表的列进行排序的功能 (RS48256)。
- 创建新的地理分布式 (Active-Active) 数据库时,不再需要终端节点端口。如果提供,则系统将分配一个 (RS27632)。
其他增强功能
-
添加了一个选项,用于为主动-主动数据库的管理层运行连接运行状况检查。运行以下 REST API 命令:
GET https:/[host][:port]/v1/crdbs/<crdb_guid>/health_report
-
Added TLS handshake error messages to the DMC proxy log (RS59346).
Resolved issues
-
RS58219 - Fixes a UI error message that showed a path instead of a relevant error message.
-
RS44958 - Fixes incorrect description for the graph "incoming traffic" in Active-Active (geo-distributed) database UI Metrics.
-
RS66280 - Fixes the lexicographic SORT command on Active-Active databases (e.g. SORT mylist ALPHA
). The SORT command should only run on keys mapped to the same slot.
-
RS64575 - Fixes a bug in the replication between primary and replica shards of a destination Active-active database in the scenario of using Replica-Of from a single to an Active-Active database, where the syncer process went down during the full sync.
-
RS65370 - Adds logic to remove old syncer entries in the cluster configuration during upgrades.
-
RS67434 - Version 6.2.10 fixes the mTLS handshake between the syncer process and the proxy (DMC), where the proxy presented a leaf certificate without its full chain to the syncer. After upgrading to 6.2.10, syncer connections using invalid certificates will break the synchronization between Active-Active instances or deployments using Replica Of when TLS is enabled. To ensure certificates are valid before upgrading do the following:
-
For Active-Active databases, run the following command from one of the clusters:
crdb-cli crdb update --crdb-guid <CRDB-GUID> --force
-
For Active-Passive (Replica Of) databases: use the admin console to verify that the destination syncer has the correct certificate for the source proxy (DMC). For details, see Configure TLS for Replica Of.
Issues resolved in build 96
-
RS67133 - An issue in Redis Enterprise Software affected replication in replica databases using RedisGraph, RediSearch, and RedisGears in specific scenarios. The problem appeared when importing an RDB file or while synchronizing target Active-Passive (ReplicaOf) databases.
This issue is fixed in Redis Enterprise Software v6.2.10-96 and RedisGraph v2.8.11. We recommend upgrading to these versions at your earliest opportunity. (Failure to upgrade can lead to data loss.)
Once the upgrades are complete, secondary shards might need to be restarted. You can use rlutil
to restart secondary shards:
rlutil redis_restart redis=<shard-id1>,<shard-id2>,...
Issues resolved in build 100
- RS74171 - A new command was added as part of Redis 6.2: XAUTOCLAIM. When used in an Active-Active configuration, this command may cause Redis shards to crash, potentially resulting in data loss. The issue is fixed in Redis Enterprise Software version 6.2.12. Additionally, we recommend enabling AOF persistence for all Active-Active configurations.
Issues resolved in build 121
- RS68668, RS72082 - Improvements for internode encryption certification rotation
- RS72304 - Avoid starting a master shard when both master and replica shards crash and the replica did not finish recovery
- RS74469 - Fix for some Redis Active-Active + Redis Streams scenarios that could lead to shard crash during backup; failure to backup
Issues resolved in build 129
- RS77003 - Add grace time to job scheduler to allow certificate rotation in case of failure due to scheduling conflicts.
- RS71112 - Update validation during db configuration to not fail due to ports associated with nodes that are no longer in the cluster. This was done to allow db configuration during adding and removing nodes as part of load balancing.
- RS78486 - Fix known issue from 6.2.10 build 100 - When using rladmin tune db to change the replica buffer size, the command appears to succeed, but the change does not take effect.
Known limitations
-
RS81463 - A shard may crash when resharding an Active-Active database with Auto Tiering . Specifically, the shard will crash when volatile keys or Active-Active tombstone keys reside in Flash memory.
-
RS78364 - When using rladmin tune db
to change the replica buffer size, the command appears to succeed, but the change does not take effect. This issue was introduced in build 100; it will be fixed in a future build of Redis Enterprise Software v6.2.10 and in the next release (v6.2.12).
-
RS63258 - Redis Enterprise Software is not currently supported on RHEL 8 with FIPS enabled.
FIPS changes system-generated keys, which can limit secure access to the cluster or the admin console via port 8443.
-
RS63375 - RHEL 7 clusters cannot be directly upgraded to RHEL 8 when hosting databases using modules.
Due to binary differences in modules between the two operating systems, you cannot directly update RHEL 7 clusters to RHEL 8 when those clusters host databases using modules. Instead, you need to create a new cluster on RHEL 8 and then migrate existing data from your RHEL 7 cluster. This does not apply to clusters that do not use modules.
All known limitations listed in the v6.2.4 release notes have been addressed.
Installation limitations
Several Redis Enterprise Software installation reference files are installed to the directory /etc/opt/redislabs/
even if you use custom installation directories.
As a workaround to install Redis Enterprise Software without using any root directories, do the following before installing Redis Enterprise Software:
-
Create all custom, non-root directories you want to use with Redis Enterprise Software.
-
Mount /etc/opt/redislabs
to one of the custom, non-root directories.
Known issues
-
The ZRANGESTORE
command, with a special zset-max-ziplist-entries
configuration can crash Redis 6.2. See Redis repository 10767 for more details.
-
RS40641 - API requests are redirected to an internal IP in case the request arrives from a node which is not the master. To avoid this issue, use rladmin cluster config
to configure handle_redirects
or handle_metrics_redirects
.
Security
Open Source Redis Security fixes compatibility
As part of Redis commitment to security, Redis Enterprise Software implements the latest security fixes available with open source Redis. The following Open Source Redis CVE’s do not affect Redis Enterprise:
-
CVE-2021-32625 - Redis Enterprise is not impacted by the CVE that was found and fixed in open source Redis since Redis Enterprise does not implement LCS. Additional information about the open source Redis fix is on the Redis GitHub page (Redis 6.2.4, Redis 6.0.14)
-
CVE-2021-32672 - Redis Enterprise is not impacted by the CVE that was found and fixed in open source Redis because the LUA debugger is unsupported in Redis Enterprise. Additional information about the open source Redis fix is on the Redis GitHub page (Redis 6.2.6, Redis 6.0.16)
-
CVE-2021-32675 - Redis Enterprise is not impacted by the CVE that was found and fixed in open source Redis because the proxy in Redis Enterprise does not forward unauthenticated requests. Additional information about the open source Redis fix is on the Redis GitHub page (Redis 6.2.6, Redis 6.0.16)
-
CVE-2021-32762 - Redis Enterprise is not impacted by the CVE that was found and fixed in open source Redis because the memory allocator used in Redis Enterprise is not vulnerable. Additional information about the open source Redis fix is on the Redis GitHub page (Redis 6.2.6, Redis 6.0.16)
-
CVE-2021-41099 - Redis Enterprise is not impacted by the CVE that was found and fixed in open source Redis because the proto-max-bulk-len CONFIG is blocked in Redis Enterprise. Additional information about the open source Redis fix is on the Redis GitHub page (Redis 6.2.6, Redis 6.0.16)security fixes for recent CVE’s. Redis Enterprise has already included the fixes for the relevant CVE’s. Some CVE’s announced for Open Source Redis do not affect Redis Enterprise due to different and additional functionality available in Redis Enterprise that is not available in Open Source Redis.
On this page