152. Longhorn卷报错 insufficient storage；precheck new replica failed

SUSE Enterprise Storage - Longhorn- 1.7.3

One or more Longhorn volumes fall into a degraded state with the mentioned error, even though the nodes have sufficient storage space. This issue might occur in an environment where the worker node is temporarily shut down or rebooted abruptly. This article outlines the known bug and a possible workaround to mitigate the error to bring the volumes back to a healthy state. 
即使节点有足够的存储空间，一个或多个 Longhorn 卷仍会因上述错误而降级。这个问题可能发生在工作节点被临时关闭或突然重启的环境中。本文概述了已知的 bug 以及可能的解决方法，以减轻错误，使体积恢复到健康状态。

• The following error is observed for the degraded volumes in the Longhorn UI:
  在 Longhorn UI 中观察到降级体积的误差如下：

Replica Scheduling Failure
Error Message: disks are unavailable;insufficient storage;precheck new replica failed

• It is also observed that the nodes have enough disk space available to schedule replicas. 
  还观察到节点有足够的磁盘空间用于调度副本。
• Upon further investigation, the longhorn-manager pod shows below error:
  进一步调查后，长角牛经理的播报显示了以下错误：

2025-08-11T07:01:23.729057940Z time="2025-08-11T07:01:23Z" level=info msg="Cannot find a reusable failed replicas for volume pvc-xxxxxxxxx-xxxx-xxxx-xxxx-xxxxxxxxxxx" func="scheduler.(*Repli
caScheduler).CheckAndReuseFailedReplica" file="replica_scheduler.go:635"

• This is a known issue as explained here: 10210. The issue is fixed in the Longhorn version 1.8.0 and above. 
  这是一个已知的问题，如这里解释：10210。这个问题在 Longhorn 1.8.0 及以上版本中得到了修复。 
• The following workaround can be implemented to resolve the issue:
  以下变通方法可以实现以解决该问题：
  • Go to Longhorn UI >> Volumes >> Select the affected Volume >> Click on the Hamburger icon and enable “Replica Disk Soft Anti-Affinity”
    进入 Longhorn UI >> Volumes >> 选择受影响的 Volume >> 点击汉堡图标并启用“Replica Disk Soft Anti-Affinity”

• When the Longhorn node is rebooted or shut down abruptly, the replica may become faulty, and Longhorn will try to rebuild the replica on a healthy node. However, the bug mentioned above prevents the rebuild and marks the volume as degraded. 
  当 Longhorn 节点被突然重启或关闭时，副本可能会出现故障，Longhorn 会尝试在健康节点上重建副本。然而，上述的 bug 阻止了重建，并将音量标记为降级。
• The workaround of enabling Disk‑level soft anti‑affinity ensures that Longhorn allows scheduling the replica correctly on any healthy node. 
  启用磁盘级软反亲和机制的变通方法确保 Longhorn 允许在任何健康节点上正确调度副本。