Yiyang Huang

> Back to the proposal content, as there are multiple DB supported, do we need to show the DB configurations in the system settings page of the web console? I...

> I can't speak for @JinXingYoung (so please correct me if this wasn't your intention), but I think the proposal as it reads currently is reasonably clear that it's the...

LGTM, Using a common base image will make it easier to maintain. Although it may cause cyclone base image size larger.

Awesome!

Could you please submit a PR to fix this issue ?

> 最近观察到日志，有些回收检查，感觉这个有点关系 2023-06-12T03:01:22.508Z DEBUG storage/local_storage.go:494 reclaim check, last access: 2023-06-12 02:56:08.386595937 +0000 UTC, reclaim: false {"task": "a0665ca1cec6191e10683a4c11b7e9489d19172b8b132225af2208314a82008a", "peer": "10.12.34.82-961540-06d7c72a-b3bf-4b81-ad3d-7d7be412fea8", "component": "io.d7y.storage.v2.simple"} 这个是 dfdaemon GC 的日志，和这个问题无关。

> 会不会和 harbor仓库是私有的，而containerd 配置的 http://127.0.0.1:65001 这里没有添加auth的原因，导致了其他问题， 从nat流量(第一次拉镜像走nat) 和peer 能确定 除了第一次走了nat，后面确实是走了peer流量，而不是每次都走harbor，但是感觉走peer的流量速度很慢 7G镜像要拉5min pod 配置 imagepullsecrets 就行，containerd 不需要配置 auth

> > @gaofengliu0 > > 可以排查三部分: > > > > 1. 瓶颈是否在 dfdameon 配置的上传和下载限速。具体配置是 `download.totalRateLimit`、`download.perPeerRateLimit`、`upload.rateLimit`，可以参考文档: [https://d7y.io/docs/reference/configuration/dfdaemon。](https://d7y.io/docs/reference/configuration/dfdaemon%E3%80%82) > > 2. 瓶颈是否在 Peer to Peer 之间实际带宽，是不是只有 100MB/s 左右，因为 metrics 采集对于突发情况数据不一定准确，要实际用工具端到端测量。 > >...

> Do we need to attach other annotations besides identifying it as a nydus image? Currently, No.

I will follow this issue