发布日期: 2026-05-05
版本号: v1.11.2

Longhorn v1.11.2版本发布，包含多项改进和错误修复，旨在提升系统质量、韧性、稳定性和安全性。该版本修复了几个关键问题，包括解决了副本重建进度在网络不稳定时可能超过100%的错误，现已将进度上报限制为100%；引入了新设置以控制CSIStorageCapacity的报告，解决了无Longhorn磁盘的计算节点错误报告0容量从而影响调度的问题。此外，优化了longhorn-manager Pod informer缓存以降低集群内存使用。安装和升级此版本需要Kubernetes v1.25或更高版本。此版本还回溯修复了多个问题，涉及稳定性、调度、快照清理和备份检查等方面。发布后已知问题和完整的问题列表可在项目仓库中查阅。

更新内容 （中文）

Longhorn v1.11.2 发布说明

Longhorn 1.11.2 引入了多项改进和错误修复，旨在提升系统质量、弹性、稳定性和安全性。

我们欢迎反馈和贡献，以帮助持续改进 Longhorn。

关于 Longhorn 版本的术语和背景信息，请参阅 Releases。

重要修复

本版本包含若干关键的稳定性修复。

副本重建进度修复

解决了一个在不稳定网络条件下，副本重建进度可能超过 100% 的问题。进度报告现在已封顶于 100%。

更多详情，请参阅 #12949。

CSIStorageCapacity 调度增强

引入了一个新设置以控制 CSIStorageCapacity 的报告。此前，没有 Longhorn 磁盘的计算节点错误地报告了 0 容量，导致 WaitForFirstConsumer 调度失败。通过此增强，容量跟踪现在可以配置，以避免在分离的计算/存储架构中拒绝计算节点。

更多详情，请参阅 #12807。

改进

Manager 内存优化

优化了 longhorn-manager Pod 信息缓存，以减少集群范围的内存使用。

更多详情，请参阅 #12771。

安装

[!IMPORTANT] 在安装 Longhorn v1.11.2 之前，请确保您的集群运行的是 Kubernetes v1.25 或更高版本。

您可以使用多种工具安装 Longhorn，包括 Rancher、Kubectl 和 Helm。有关安装方法和要求的更多信息，请参阅 Longhorn 文档中的 快速安装。

升级

[!IMPORTANT] 在从 Longhorn v1.10.x 或 v1.11.0 升级到 v1.11.2 之前，请确保您的集群运行的是 Kubernetes v1.25 或更高版本。

[!IMPORTANT] 强烈建议 v1.11.0 版本的用户升级到 v1.11.1 或更高版本，以获取针对 longhorn-instance-manager Pod 内存泄漏问题 12575 的永久修复（针对代理连接泄漏）。

Longhorn 仅允许从受支持的版本升级。有关升级路径和程序的更多信息，请参阅 Longhorn 文档中的 升级。

发布后已知问题

有关此版本发布后识别出的问题，请参阅 Release-Known-Issues。

本版本中已解决的问题

改进

  • [BACKPORT][v1.11.2][IMPROVEMENT] 通过优化集群范围的信息缓存减少 longhorn-manager 内存使用 12819 - @hookak @roger-ryao

错误

  • [BACKPORT][v1.11.2][BUG] 测试用例 test_storage_capacity_aware_pod_scheduling 失败 13006 - @yangchiu @bachmanity1
  • [BACKPORT][v1.11.2][BUG] 副本自动平衡导致无限副本调度循环 12928 - @yangchiu @shuo-wu
  • [BACKPORT][v1.11.2][BUG] CSIStorageCapacity 为没有 Longhorn 磁盘的计算节点报告 0 容量，破坏 WaitForFirstConsumer 调度 12918 - @chriscchien @bachmanity1
  • [BACKPORT][v1.11.2][BUG] 副本重建进度可能超过 100% 12952 - @yangchiu @davidcheng0922
  • [BACKPORT][v1.11.2][BUG] 由 NFS 延迟导致的备份检查积累引起的节点耗尽 12945 - @COLDTURNIP @roger-ryao
  • [BACKPORT][v1.11.2][BUG] 当磁盘路径是 /dev/disk/by-id 软链接时，未能收集块磁盘 (AIO) 的健康数据 12911 - @yangchiu @hookak
  • [BACKPORT][v1.11.2][BUG] 在备份后的预期自动清理期间发出 “snapshot becomes not ready to use” 警告事件 12856 - @EpochBoy @yangchiu

稳定性

  • [BACKPORT][v1.11.1][BUG] 卷指标收集器中可能的 NEP 12733 - @derekbit @chriscchien

贡献者

  • @COLDTURNIP
  • @bachmanity1
  • @chriscchien
  • @davidcheng0922
  • @derekbit
  • @EpochBoy
  • @github-actions[bot]
  • @hookak
  • @innobead
  • @roger-ryao
  • @shuo-wu
  • @yangchiu
  • @sushant-suse
  • @rebeccazzzz
  • @forbesguthrie

更新内容 （原始）

Longhorn v1.11.2 Release Notes

Longhorn 1.11.2 introduces several improvements and bug fixes that are intended to improve system quality, resilience, stability and security.

We welcome feedback and contributions to help continuously improve Longhorn.

For terminology and context on Longhorn releases, see Releases.

Important Fixes

This release includes several critical stability fixes.

Replica rebuild progress fix

Resolved an issue where replica rebuild progress could exceed 100% under unstable network conditions. Progress reporting is now capped at 100%.

For more details, see #12949.

CSIStorageCapacity scheduling enhancement

Introduced a new setting to control CSIStorageCapacity reporting. Previously, compute nodes without Longhorn disks incorrectly reported 0 capacity, breaking WaitForFirstConsumer scheduling. With this enhancement, capacity tracking can be configured to avoid rejecting compute nodes in separated compute/storage architectures.

For more details, see #12807.

Improvement

Manager memory optimization

Optimized longhorn‑manager Pod informer caching to reduce cluster‑wide memory usage.

For more details, see #12771.

Installation

[!IMPORTANT] Ensure that your cluster is running Kubernetes v1.25 or later before installing Longhorn v1.11.2.

You can install Longhorn using a variety of tools, including Rancher, Kubectl, and Helm. For more information about installation methods and requirements, see Quick Installation in the Longhorn documentation.

Upgrade

[!IMPORTANT] Ensure that your cluster is running Kubernetes v1.25 or later before upgrading from Longhorn v1.10.x or v1.11.0 to v1.11.2.

[!IMPORTANT] Users on v1.11.0 who experienced the memory leaks of longhorn-instance-manager pods 12575 are highly encouraged to upgrade to v1.11.1 or later to receive the permanent fix for the proxy connection leaks.

Longhorn only allows upgrades from supported versions. For more information about upgrade paths and procedures, see Upgrade in the Longhorn documentation.

Post-Release Known Issues

For information about issues identified after this release, see Release-Known-Issues.

Resolved Issues in this release

Improvement

  • [BACKPORT][v1.11.2][IMPROVEMENT] Reduce longhorn-manager memory usage by optimizing cluster-wide informer caching 12819 - @hookak @roger-ryao

Bug

  • [BACKPORT][v1.11.2][BUG] Test case test_storage_capacity_aware_pod_scheduling fails 13006 - @yangchiu @bachmanity1
  • [BACKPORT][v1.11.2][BUG] Replica Auto-Balance Causes Infinite Replica Scheduling Loop 12928 - @yangchiu @shuo-wu
  • [BACKPORT][v1.11.2][BUG] CSIStorageCapacity reports 0 for compute nodes without Longhorn disks, breaking WaitForFirstConsumer scheduling 12918 - @chriscchien @bachmanity1
  • [BACKPORT][v1.11.2][BUG] Replica rebuild progress can go over 100% 12952 - @yangchiu @davidcheng0922
  • [BACKPORT][v1.11.2][BUG] Node exhaustion caused by backup inspect buildup induced due to NFS latency 12945 - @COLDTURNIP @roger-ryao
  • [BACKPORT][v1.11.2][BUG] Failed to collect health data for block disk (AIO) when disk path is a /dev/disk/by-id symlink 12911 - @yangchiu @hookak
  • [BACKPORT][v1.11.2][BUG] “snapshot becomes not ready to use” Warning events emitted during expected auto-cleanup after backup 12856 - @EpochBoy @yangchiu

Stability

  • [BACKPORT][v1.11.1][BUG] Potential NEP in Volume Metrics Collector 12733 - @derekbit @chriscchien

Contributors

  • @COLDTURNIP
  • @bachmanity1
  • @chriscchien
  • @davidcheng0922
  • @derekbit
  • @EpochBoy
  • @github-actions[bot]
  • @hookak
  • @innobead
  • @roger-ryao
  • @shuo-wu
  • @yangchiu
  • @sushant-suse
  • @rebeccazzzz
  • @forbesguthrie

下载链接