108. S3 备份在 Rancher 配置的 RKE2 或 K3s 集群中失败，且显示“未能测试桶的存在：HEAD Forbidden”

A Rancher-provisioned RKE2 or K3s cluster, with etcd snapshots configured to an S3 bucket, and a HTTP Proxy configuration
一个由 Rancher 配置的 RKE2 或 K3s 集群，配置了 etcd 快照到 S3 桶，并配置了 HTTP 代理

The automated etcd snapshots for a Rancher-provisioned RKE2 or K3s cluster to an S3 bucket are failing. The error message listed next to the Failed snapshots is of the following format:
Rancher 配置的 RKE2 或 K3s 集群到 S3 桶的自动 etcd 快照失败了。失败快照旁边列出的错误信息格式如下：

failed to initialize S3 client: failed to test for existence of bucket etcd-backups: Head "https://s3.example.com/etcd-backups/": Forbidden

In addition, in Rancher < v2.11 you may experience some UI slowness as a result of the repeated failing snapshots in the affected cluster(s).
此外，在 Rancher < v2.11 中，由于受影响集群快照反复失败，你可能会遇到界面变慢的情况。

The recommended approach to resolve this is to double check the proxy configuration of the downstream RKE2/K3s cluster, specifically theNO_PROXYsetting. Ensure this contains an entry that matches the full hostname of the S3 endpoint, e.g. s3.example.com in the example above.
建议的解决方法是再次检查下游 RKE2/K3s 集群的代理配置，特别是NO_PROXY设置。确保该条目与 S3 端点的完整主机名相匹配，例如上述示例中的 s3.example.com。

The cause for this case is determined to be a missing entry in the affected cluster's NO_PROXY environment variable, which can result in a "HEAD <S3-ENDPOINT:PORT> Forbidden" error, as a result of S3 requests being incorrectly sent to the HTTP Proxy.
导致此情况的原因被确定为受影响集群 NO_PROXY 环境变量中缺少一个条目，可能导致“HEAD <S3-ENDPOINT：PORT> Forbidden”错误，因为 S3 请求错误地发送到 HTTP 代理。