Redis Release Notes

Redis Enterprise Software release notes 7.4.x

New Cluster Manager UI enhancements, including Active-Active database management. Full TLS 1.3 support. Automatic recovery configuration. Full IPv6 support, including for internal traffic. Maintenance mode enhancements. Module management enhancements. RHEL 9 support.

Highlights

This version offers:

• New Cluster Manager UI enhancements, including Active-Active database management
• Full TLS 1.3 support
• Automatic recovery configuration
• Full IPv6 support, including for internal traffic
• Maintenance mode enhancements
• Module management enhancements
• RHEL 7 and Oracle Linux 7, which were previously deprecated, are no longer supported
• RHEL 9 support

Detailed release notes

7.4.6-293 (October 2026): Ubuntu 18.04 is no longer supported. Internal fixes and improvements.

Redis Software 7.4 reaches end of life on November 30, 2026. After that date, 7.4.x clusters no longer receive security patches, bug fixes, or maintenance releases, and support for 7.4-specific issues is limited per your subscription agreement.

Plan your upgrade to a supported version. See the supported upgrade paths and the Redis Software product lifecycle.

The end-of-life policy for Redis Enterprise Software versions 6.2 and later has been extended to 24 months after the formal release of the subsequent major version.

Redis Enterprise Software versions 6.2.4 and 6.2.8 do not support direct upgrades beyond version 7.4.x. Versions 6.2.10, 6.2.12, and 6.2.18 are part of the upgrade path. To upgrade from 6.2.4 or 6.2.8 to versions later than 7.4.x, an intermediate upgrade is required.

The next major Redis Enterprise Software release will still bundle Redis database version 6.2 and allow database upgrades from Redis database version 6.2 to 7.x.

Deprecations

The replica HA cluster policy is deprecated.

The maintenance mode option keep_slave_shards is deprecated in rladmin and the REST API. Use evict_ha_replica and evict_active_active_replica instead.

The /v1/debuginfo REST API paths are deprecated. Use the new paths /v1/cluster/debuginfo, /v1/nodes/debuginfo, and /v1/bdbs/debuginfo instead.

The legacy UI is deprecated in favor of the new Cluster Manager UI and will be removed in a future release.

Redis database version 6.0 is deprecated as of Redis Enterprise Software version 7.4.2 and will be removed in a future release.

The triggers and functions (RedisGears) preview has been discontinued. Commands such as TFCALL, TFCALLASYNC, and TFUNCTION will be deprecated and will return error messages. Any JavaScript functions stored in Redis will be removed. JavaScript-based triggers will be blocked. Lua functions and scripts will not be affected.

Redis has announced the end of life for RedisGraph. Redis will continue to support all RedisGraph customers, including releasing patch versions until January 31, 2025.

RHEL 7 and Oracle Linux 7 are no longer supported. The RC4 encryption cipher, the 3DES encryption cipher, and TLS 1.0 and TLS 1.1 connections are no longer supported.

Upcoming changes

Starting with the next major version, Redis Enterprise Software containers with the image tag x.y.z-build will be based on RHEL instead of Ubuntu. To use Ubuntu-based images after this change, specify the operating system suffix in the image tag, such as 7.4.2-216.focal instead of 7.4.2-216.

Known issues

RS131972: Creating an ACL that contains a line break in the Cluster Manager UI can cause shard migration to fail due to ACL errors.

RS61676: Full chain certificate update fails if any certificate in the chain does not have a Common Name (CN).

RS119958: The debuginfo script fails with the error /bin/tar: Argument list too long if there are too many RocksDB log files. This issue only affects clusters with Auto Tiering.

RS122570: REST API POST /crdbs responds with a confusing error message if the cluster does not have the requested CRDB-compatible module that complies with the requested featureset. This issue was fixed in Redis Enterprise Software version 7.4.2-126.

RS123142: In an Active-Active setup with at least three participating clusters, removing and re-adding a cluster after removing older clusters without re-adding them can cause missing keys and potentially lead to data loss or data inconsistency. This issue was fixed in Redis Enterprise Software version 7.4.2-169.

Known limitations

The following legacy UI features are not yet available in the new Cluster Manager UI: remove a node, purge an Active-Active instance, and search and export the log.

Due to an OpenSSL 1.1 compatibility issue between modules and clusters, Redis Enterprise Software version 7.4.2-54 is not fully supported on Amazon Linux 2 clusters with databases that use RedisGears, RediSearch, or RedisTimeSeries.

You cannot upgrade from a prior RHEL version to RHEL 9 if the Redis Enterprise cluster contains a RedisGraph module.

You cannot create Active-Active databases that use Redis version 6.0 or 6.2 with modules. Databases that use Redis version 7.2 do not have this limitation.

When you install Redis Enterprise Software version 7.4.2 on RHEL 9, firewalld configuration can fail due to file permissions. This limitation was fixed in Redis Enterprise Software version 7.4.2-129.

Redis Software release notes 8.2.x

Database-scoped management roles. Data path command auditing.

Redis Software version 8.2 is now available!

Highlights

This version offers:

• Database-scoped management roles
• Data path command auditing
• FIPS 140-3 support

Detailed release notes

For more detailed release notes, select a build version from the following table:

8.2.0-125 (October 2026)

Bug fixes for upgrades from Redis Software 7.2.x and 7.4.x, cluster management after a primary node change, node removal during upgrades, Active-Active proxy recovery, DMC proxy stability, scaling and resharding, ACL validation, cluster recovery, Google Cloud Storage backup configuration, and signing in with passwords that contain a colon. Internal fixes and improvements.

Redis 8.6.2, 8.4.0, 8.2.1, 8.0.2, 7.4.3, 7.2.7, 6.2.13

8.2.0-78 (September 2026)

Bug fixes for minor version upgrades, shard placement, LDAP client authentication, database connectivity, DMC proxy connection cleanup, internode encryption certificates, Auto Tiering database resizing, crdb-cli certificate-based authentication, metrics export, module configuration, and support package collection. Internal fixes and improvements.

Redis 8.6.2, 8.4.0, 8.2.1, 8.0.2, 7.4.3, 7.2.7, 6.2.13

8.2.0-46 (August 2026)

Bug fixes for internal certificate rotation, rolling upgrades, the Cluster Manager UI, shard rebalancing, and Active-Active synchronization. Internal fixes and improvements.

Redis 8.6.2, 8.4.0, 8.2.1, 8.0.2, 7.4.3, 7.2.7, 6.2.13

8.2.0-25 (July 2026)

Database-scoped management roles. Data path command auditing.

Redis 8.6.2, 8.4.0, 8.2.1, 8.0.2, 7.4.3, 7.2.7, 6.2.13

Breaking changes

Redis Software version 8.2.0 introduces the following breaking changes:

• Local import from Redis Software system directories is no longer supported. To improve security, database imports that use the mount_point source can no longer read files from Redis Software system directories, including /opt/redislabs, /var/opt/redislabs, /etc/opt/redislabs, and any of their subdirectories.
  • Import operations that reference files in these directories now fail.
  • Store import files in a non-system location accessible to Redis Software before you start the import.
• smtp_password is now a write-only field in the Cluster API. To improve security, the API no longer returns the SMTP password in responses to GET /v1/cluster, and the Cluster Manager UI no longer reads or displays the existing password.
  • Applications or scripts that previously expected smtp_password to be returned by the Cluster API no longer receive this field.
  • When you update SMTP settings, the existing password cannot be retrieved. To change the SMTP password, provide a new password as part of the update request.
• Custom module upload is disabled by default. To improve security, custom (user-defined) module uploads are now disabled by default. A new cluster setting, allow_modules_upload, controls access to the custom module upload APIs. When this setting is disabled (the default), requests to POST /v2/modules/user-defined and POST /v2/local/modules/user-defined/artifacts return 403 modules_upload_disabled, regardless of user role.
  • To enable custom module uploads, set allow_modules_upload to true with a PUT /v1/cluster request.
  • Applications or automation that call the custom module upload APIs directly receive a 403 modules_upload_disabled response until the setting is enabled.
  • Existing functionality for listing, deleting, and loading modules is unchanged.

Reserved ports

Make sure the following ports are open before upgrading Redis Software.

The following port was added as a reserved port in Redis Software version 8.0.18; however, it is optional instead of reserved in Redis Software version 8.0.20-44:

3357 — reconciliation_tree_grpc — Internal communication

Ports reserved as of Redis Software version 7.22.0:

3346 — cluster_api_internal — Cluster API internal port
3351 — cluster_watchdog_grpc_api — Cluster watchdog now supports gRPC
3352 — grpc_service_mesh — gRPC communication between nodes
3353 — local_grpc_service_mesh — Local gRPC services
3354 — grpc_gossip_envoy — gRPC gossip protocol communication between nodes
3355 — authentication_service — Authentication service internal port

Ports reserved as of Redis Software version 7.8.2:

3347 — cert_exporter — Reports cluster certificate metrics
3348 — process_exporter — Reports process metrics for DMC and Redis processes
3349 — cluster_wd_exporter — Reports cluster watchdog metrics
3350 — db_controller — Internode communication
9091 — node_exporter — Reports host node metrics related to CPU, memory, disk, and more
9125 — statsd_exporter — Reports push metrics related to the DMC and syncer, and some cluster and node metrics

Deprecations

API deprecations

• The db_conns_auditing cluster policy and database configuration field is deprecated. Use the audit_settings.audit_mode field instead. Setting db_conns_auditing to true is equivalent to setting audit_mode to connection. See Audit events for the current auditing configuration.
• The gradual_src_mode, gradual_sync_mode, and gradual_sync_max_shards_per_source parameters are deprecated. Use gradual_sync_policy instead. You can set it with rladmin tune db or the REST API.
• The old_password field in POST /v1/users/password and PUT /v1/users/password requests is deprecated and ignored. It was optional and had no effect when adding or resetting a password, so you can remove it from these requests. old_password is still required for DELETE /v1/users/password, which uses it to identify the password to delete.
• The following REST API fields, deprecated since Redis Software version 6.4.2, are removed in this release:
  • use_ipv6 (cluster object) — use use_external_ipv6 instead.
  • redis_cleanup_job_settings (job scheduler object) — use persistence_cleanup_scan_interval instead.

Internal monitoring and v1 Prometheus metrics deprecation

The existing internal monitoring engine is deprecated. We recommend transitioning to the new metrics stream engine for improved performance, enhanced integration capabilities, and modernized metrics streaming.

V1 Prometheus metrics are deprecated but still available. To transition to the new metrics stream engine, either migrate your existing dashboards using this guide or use new preconfigured dashboards.

As part of the transition to the metrics stream engine, some internal cluster manager alerts were deprecated in favor of external monitoring solutions. See the alerts transition plan for guidance.

Supported platforms

The following table provides a snapshot of supported platforms as of this Redis Software release. See the supported platforms reference for more details about operating system compatibility.

✅ Supported – The platform is supported for this version of Redis Software and Redis Stack modules.

⚠️ Deprecation warning – The platform is still supported for this version of Redis Software, but support will be removed in a future release.

Release date: Jul 2026

Supported platforms include RHEL 9 and compatible distributions, RHEL 9 FIPS mode, RHEL 8 and compatible distributions, Ubuntu 22.04, Ubuntu 20.04, Amazon Linux 2023, Kubernetes, and Docker.

Known issues

• RS155734: Endpoint availability metrics do not work as expected due to a calculation error.

Known limitations

Redis Search query failures during rolling upgrades across the 8.4 version boundary

Redis Search queries can fail during a rolling upgrade when a cluster contains shards running Redis versions earlier than 8.4 and shards running Redis version 8.4 or later, due to an internal protocol change introduced in version 8.4. This issue affects only clusters where parallel_shards_upgrade has been changed from its default value of 0. If both conditions apply, expect Redis Search downtime until all nodes are upgraded.

Trim ACKED not supported for Active-Active 8.4 databases

For Active-Active databases running Redis database version 8.4, the ACKED option is not supported for trimming commands.

Rolling upgrade limitation for clusters with custom or deprecated modules

Due to module handling changes introduced in Redis Software version 8.0, upgrading a cluster that contains custom or deprecated modules, such as RedisGraph and RedisGears v2, can become stuck when adding a new node to the cluster during a rolling upgrade.

Module commands limitation during Active-Active database upgrades to Redis 8.0

When upgrading an Active-Active database to Redis version 8.0, you cannot use module commands until all Active-Active database instances have been upgraded. Currently, these commands are not blocked automatically.

Redis 8.0 database cannot be created with flash

You cannot create a Redis 8.0 database with flash storage enabled. Create a Redis 8.0 database with RAM-only storage instead, or use Redis 8.2 for flash-enabled (Redis Flex) databases.

New Cluster Manager UI limitations

The following legacy UI features are not yet available in the new Cluster Manager UI:

• Purge an Active-Active instance. Use crdb-cli crdb purge-instance instead.
• Search and export the log.

SCAN results can be inconsistent during a rolling upgrade in OSS Cluster API mode

During a rolling upgrade with Smart client handoffs (SCH) enabled, when some slot's shards have migrated to a different node but are still reachable from their original node, results of the SCAN command might be inconsistent: it might show some of the keys, it might show keys that are not reachable from that node, or it might show no keys. When the slot migration is done, this issue will fix itself.

Note

Low Risk

Changes are limited to the Tcl test harness timeout path; no production server or runtime behavior is affected.

Overview

When the suite hits --timeout (no client progress), it no longer tears down immediately after printing each client’s last task. The test server collects diagnostics first, then kills clients and servers as before.

Server evidence: Still-running Redis instances from ::active_servers get SIGCONT (if stopped) and SIGSEGV so they write a full crash report even when the event loop is wedged. Logs are located under tests/tmp by pid, then dump_crash_report prints the tail around REDIS BUG REPORT START (or the last 256KB if there is no report). is_running treats zombies as dead so waits don’t hang on unreaped children.

Client evidence: Clients send their OS pid on ready and optionally advertise sigusr1-trace when Tclx is available. After server dumps, the server waits briefly for natural exception/err unwinds, then sends SIGUSR1 to remaining clients so Tclx turns it into a stack trace. ::in_timeout_report suppresses re-entrant timeout cron, avoids fatal handling of those packets, and fixes read_from_test_client when a client disconnects mid-report (invalid length no longer spins or crashes the handler).

Order: Server crash collection runs before client stack traces so unblocking a client doesn’t tear down servers before their reports are captured.

Reviewed by Cursor Bugbot for commit 5927e7e. Bugbot is set up for automated code reviews on this repo. Configure here.

Co-authored-by: Claude Opus 5.5 (1M context) [email protected]

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

• #15673 Commands queued in a transaction could still access keys whose ACL permissions were revoked before the transaction was executed
• #15722 The cluster bus protocol has no authentication of its own unless tls-cluster is enabled, so any host able to reach a node's bus port could join the cluster and threaten it. A cluster node now warns at startup when its bus port is left unauthenticated, and the new cluster-bus-port-protected-mode option (default no) makes refusing to run in that state an explicit choice: set it to yes and the node starts only when tls-cluster authenticates the bus
• TimeSeries: Prevented Redis from crashing when adding samples to a compressed Time Series key restored from a malformed RDB payload
• RedisSearch: KNN queries on indexes with very long vector field names could cause the server to crash
• Vector Sets: Deeply nested JSON used in Vector Set queries could cause the server to crash

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

• #15673 Commands queued in a transaction could still access keys whose ACL permissions were revoked before the transaction was executed
• #15722 The cluster bus protocol has no authentication of its own unless tls-cluster is enabled, so any host able to reach a node's bus port could join the cluster and threaten it. A cluster node now warns at startup when its bus port is left unauthenticated, and the new cluster-bus-port-protected-mode option (default no) makes refusing to run in that state an explicit choice: set it to yes and the node starts only when tls-cluster authenticates the bus
• TimeSeries: Prevented Redis from crashing when adding samples to a compressed Time Series key restored from a malformed RDB payload
• Vector Sets: Deeply nested JSON used in Vector Set queries could cause the server to crash

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

• #15673 Commands queued in a transaction could still access keys whose ACL permissions were revoked before the transaction was executed
• #15722 The cluster bus protocol has no authentication of its own unless tls-cluster is enabled, so any host able to reach a node's bus port could join the cluster and threaten it. A cluster node now warns at startup when its bus port is left unauthenticated, and the new cluster-bus-port-protected-mode option (default no) makes refusing to run in that state an explicit choice: set it to yes and the node starts only when tls-cluster authenticates the bus
• TimeSeries: Prevented Redis from crashing when adding samples to a compressed Time Series key restored from a malformed RDB payload
• Vector Sets: Deeply nested JSON used in Vector Set queries could cause the server to crash

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

• #15673 Commands queued in a transaction could still access keys whose ACL permissions were revoked before the transaction was executed
• #15722 The cluster bus protocol has no authentication of its own unless tls-cluster is enabled, so any host able to reach a node's bus port could join the cluster and threaten it. A cluster node now warns at startup when its bus port is left unauthenticated, and the new cluster-bus-port-protected-mode option (default no) makes refusing to run in that state an explicit choice: set it to yes and the node starts only when tls-cluster authenticates the bus
• TimeSeries: Prevented Redis from crashing when adding samples to a compressed Time Series key restored from a malformed RDB payload
• Vector Sets: Deeply nested JSON used in Vector Set queries could cause the server to crash

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

#15722 The cluster bus protocol has no authentication of its own unless tls-cluster is enabled, so any host able to reach a node's bus port could join the cluster and threaten it. A cluster node now warns at startup when its bus port is left unauthenticated, and the new cluster-bus-port-protected-mode option (default no) makes refusing to run in that state an explicit choice: set it to yes and the node starts only when tls-cluster authenticates the bus

TimeSeries: Prevented Redis from crashing when adding samples to a compressed Time Series key restored from a malformed RDB payload

Vector Sets: Deeply nested JSON used in Vector Set queries could cause the server to crash

Redis Software version 7.8 is now available!

Highlights

This version offers:

• Redis Community Edition 7.4 features
• Hash field expiration
• Client-side caching support
• Metrics stream engine preview
• New APIs to check database availability, rebalance shards, fail over shards, and control database traffic
• Cluster Manager UI enhancements for node actions, database tags, and database configuration
• User manager role
• Log rotation based on both size and time
• Module management enhancements
• Configurable minimum password length
• Configurable license expiration alert threshold

Detailed release notes

For more detailed release notes, select a build version from the following table:

7.8.6-315 (September 2026): Bug fix for a memory leak in the Active-Active database syncer. Added scoping for shards and database endpoints. Internal fixes and improvements.

Redis 7.8.6-303 (August 2026): Internal fixes and improvements.

Redis 7.8.6-286 (July 2026): Internal fixes and improvements.

Redis 7.8.6-263 (May 2026): Internal fixes and improvements.

Redis 7.8.6-260 (April 2026): Internal fixes and improvements.

Redis 7.8.6-256 (March 2026): Internal fixes and improvements.

Redis 7.8.6-253 (February 2026): Bug fixes for database recovery and modifying databases in the Cluster Manager UI. Internal fixes and improvements.

Redis 7.8.6-207 (September 2025): Enhanced timeout diagnostics and logging when restarting shards. Bug fix for viewing shard memory usage metrics in the Cluster Manager UI. Internal fixes and improvements.

Redis 7.8.6-152 (August 2025): Bug fixes for Active-Active syncing, event log entries for configuration changes made by LDAP users, LDAP mapping visibility in the Cluster Manager UI, database max_connections setting, and memory leaks when using Active-Active databases in Kubernetes.

Redis 7.8.6-119 (June 2025): Bug fixes for user password management REST API requests, maintenance mode snapshot cleanup, and rlec_supervisor file permissions.

Redis 7.8.6-95 (May 2025): Performance improvements for the Cluster Manager UI's nodes screen. Bug fix for slowness and intermittent failures in the Cluster Manager UI, rladmin, and rlcheck after upgrading large clusters to version 7.8.4. Bug fix for an issue that could cause inconsistent memory settings across shards and out-of-memory errors. Bug fix for an issue where certain processes might not terminate as expected during cluster upgrades due to timeouts.

Redis 7.8.6-60 (April 2025): Bug fixes for an LDAP connection issue caused by setting LDAP bind credentials that include < in the Cluster Manager UI and an issue where migrated shards can appear stuck on multiple nodes. Deprecated required_version in the bootstrap cluster API.

Redis 7.8.6-36 (April 2025): Bug fixes, including a fix for an issue where new workers might use an outdated certificate after a certificate update.

Redis 7.8.6-13 (March 2025): Configurable hashing algorithm for user passwords. Dry run option to validate REST API requests for users and roles.

Redis 7.8.4-95 (March 2025): Bug fixes for server restart failures during upgrades, connectivity checks for optional services, swapped shard status and watchdog status values in rladmin, process exporter crashes, and a syncer issue for Active-Active databases that have the OSS Cluster API enabled.

Redis 7.8.4-66 (February 2025): Ubuntu 22 support. Bug fixes for crashes, unexpected restarts, and incorrect type for the automatically calculated replica buffer.

Redis 7.8.4-18 (December 2024): Certificate-based authentication for the REST API. Dry run option to validate Redis ACL REST API requests.

Redis 7.8.2-60 (December 2024): Bug fixes, including a fix for a RHEL 8 OpenSSL dependency issue that caused frequent restarts of the cm_server and prevented access to the Redis Software Cluster Manager UI after upgrading to Redis Software version 7.8.2-34.

Redis 7.8.2-34 (November 2024): Redis Community Edition 7.4 features. Hash field expiration. Client-side caching support. Metrics stream engine preview. New APIs to check database availability, rebalance shards, fail over shards, and control database traffic. Cluster Manager UI enhancements for node actions, database tags, and database configuration. User manager role. Log rotation based on both size and time. Module management enhancements. Configurable minimum password length. Configurable license expiration alert threshold.

Update urgency: SECURITY: There are security fixes in the release.

Security fixes

• (CVE-2026-62356) Miscalculated buffer size in CMSketch RDB loading may lead to heap OOB write
• Out-of-bounds access in TopK heap cleanup path (MOD-15410)
• Use-after-free in the TLS pending-data list when a command closes another pending connection
• A malicious RDB payload with an out-of-range SLOT_INFO slot id causes memory corruption during RDB loading, which may lead to Remote Code Execution
• Vector Sets: missing node level validation when loading a vector set from RDB may lead to out-of-bounds access
• Vector Sets: use-after-free when VREM mutates the HNSW graph while background VSIM threads are still running
• Vector Sets: a negative hnsw_search() return was treated as a huge unsigned count, reading past the end of the result arrays
• TLS client certificate authentication bypass: a Common Name containing an embedded NUL byte was truncated, allowing a client to authenticate as another (possibly privileged) ACL user
• #15594 Use-after-free in the blocked-client list when reprocessing a command evicts another client blocked on the same key