From abba58e30b5bce9034df673301f64e765f1bd188 Mon Sep 17 00:00:00 2001 From: Rachel Elledge Date: Wed, 29 May 2024 14:21:30 -0500 Subject: [PATCH] DOC-3486 Feedback updates for enhancements and resolved issues --- .../rs-7-4-2-releases/rs-7-4-2-169.md | 22 +++++++++---------- 1 file changed, 10 insertions(+), 12 deletions(-) diff --git a/content/operate/rs/release-notes/rs-7-4-2-releases/rs-7-4-2-169.md b/content/operate/rs/release-notes/rs-7-4-2-releases/rs-7-4-2-169.md index e52231299..756e8579f 100644 --- a/content/operate/rs/release-notes/rs-7-4-2-releases/rs-7-4-2-169.md +++ b/content/operate/rs/release-notes/rs-7-4-2-releases/rs-7-4-2-169.md @@ -6,7 +6,7 @@ categories: - operate - rs compatibleOSSVersion: Redis 7.2.0 -description: cnm_http_workers and cnm_http_max_threads_per_worker are configurable. CCS CPU usage improvements. +description: cnm_http_workers and cnm_http_max_threads_per_worker are configurable. Reduced CCS CPU usage. linkTitle: 7.4.2-169 (May 2024) weight: 68 --- @@ -19,7 +19,7 @@ This version offers: - Configuration options for `cnm_http` workers and threads -- CCS CPU usage improvements +- Reduced CCS CPU usage - Bug fixes @@ -27,9 +27,9 @@ This version offers: ### Enhancements -- Configure `cnm_http_workers` and `cnm_http_max_threads_per_worker` using a [`PUT /v1/cluster`]({{}}) REST API request. +- Allow configuring `cnm_http_workers` and `cnm_http_max_threads_per_worker` using a [`PUT /v1/cluster`]({{}}) REST API request. You can update these settings if the API load is higher than the cluster can handle by default. Configuring more workers than the default will consume more memory. -- Improved Cluster Configuration Store (CCS) CPU usage +- Reduced CCS CPU usage, following the addition of the internal caching mechanism in the resource management calculations. #### Redis module feature sets @@ -61,17 +61,15 @@ Bundled Redis modules compatible with Redis database versions 6.0 and 6.2: ### Resolved issues -- RS122221: `redis-server` creates PID file only after binding to a socket. - - RS123237: CRDT syncer prevents potential data mismatches or data loss due to missing keys when removing and re-adding a cluster in an Active-Active setup with at least three participating clusters after removing older clusters without re-adding them. -- RS93876: Optimized resharding order. +- RS93876: When resharding is performed with `shards_blueprint`, improved the order of resharding to avoid stuck state machines due to low memory. - RS119350: Avoid removing roles bound to LDAP mappings during cluster upgrade even if the roles aren't assigned to any users. -- RS123561: Enclosed `$PORT_DNS` with double quotes so the variable is expanded when the install script checks if the port is free. +- RS123561: Better identification for the occupied DNS port to ensure successful installation. -- RS122365: Improved platform and module readiness checks. A module is ready if all platforms in the cluster are ready. +- RS122365: Improved platform and module readiness checks so they ignore operating systems that are no longer part of the cluster. - RS96400: DMC management threads properly terminate instead of maintaining a blocking connection if they can't reach the CCS after a predefined time. @@ -83,13 +81,13 @@ Bundled Redis modules compatible with Redis database versions 6.0 and 6.2: - RS115382: Updated DMC to disconnect the client with an error if the token size exceeds `MAX_KEY_BULK_SIZE`. -- RS110798: Made several improvements to reduce `cluster_wd` request latency. +- RS110798: Made several improvements to reduce `cluster_wd` request latency, fixing communication errors on large clusters. -- RS119705: Upgrade CCS schema before restarting services to prevent possible DMC downtime. +- RS119705: On in-place upgrades, minimize DMC downtime by performing CCS schema upgrade before services are restarted. - RS117224: Avoid RAM eviction during scripts. -- RS122062: Use `os_family` instead of `os_name` when creating module platforms during cluster upgrade. +- RS122062: Use `os_family` instead of `os_name` when creating module platforms during cluster upgrade, fixing join requests for operating systems like CentOS and Oracle Linux. ## Version changes