对使用率低而占用空间大的镜像进行统计并考虑删除

[z4yx]

从访问日志看部分镜像（如rhel的源码仓库）使用率极低，且占用空间大。我们考虑根据日志找出这类镜像，经讨论后删除，以便给更多新镜像留出空间。

[jiegec]

最近三个月的统计数据（删掉了访问量>10k的镜像）：

requests to crates.io-index.git: 8 size=["unknown", "unknown"]
requests to qemu.git: 8 size=["unknown", "unknown"]
requests to binutils-gdb.git: 12 size=["unknown", "unknown"]
requests to openthos-src: 12 size=["unknown", "unknown"]
requests to glibc.git: 16 size=["unknown", "unknown"]
requests to linux-stable.git: 26 size=["unknown", "unknown"]
requests to linux.git: 38 size=["unknown", "unknown"]
requests to chromiumos: 59 size=["unknown", "unknown"]
requests to lineageOS: 59 size=["unknown", "unknown"]
requests to llvm: 80 size=["unknown", "unknown"]
requests to packagist: 287 size=["unknown"]
requests to linux-next.git: 302 size=["unknown", "unknown"]
requests to linux-firmware.git: 501 size=["unknown", "unknown"]
requests to homebrew: 670 size=["unknown", "unknown"]
requests to gcc.git: 1205 size=["unknown", "unknown"]
requests to chakra-releases: 1774 size=["68.99G", "68.99G"]
requests to bananian: 1887 size=["unknown", "2.78G"]
requests to dotdeb: 2146 size=["25.22G", "25.22G"]
requests to slackwarearm: 2233 size=["28.84G", "28.84G"]
requests to linuxbrew-bottles: 2361 size=["unknown", "23.2G"]
requests to bjlx: 2486 size=["3.63G", "3.63G"]
requests to nix: 2487 size=["unknown", "1.49G"]
requests to pybombs: 2723 size=["unknown", "unknown"]
requests to openresty: 2899 size=["unknown", "5.17G"]
requests to elvish: 3050 size=["1.65G", "1.65G"]
requests to loongson: 3796 size=["16.87G", "16.87G"]
requests to gnu-alpha: 4351 size=["30.63G", "30.63G"]
requests to dell: 5362 size=["933.77G", "874G"]
requests to AOSP: 6266 size=["unknown", "unknown"]
requests to gentoo-portage-prefix: 7911 size=["206.43M", "206.43M"]
requests to clickhouse: 9021 size=["575.86G", "575.86G"]
requests to kali-security: 9242 size=["128K"]

去掉最近新加的镜像后：

requests to openthos-src: 12 size=["unknown", "unknown"]
requests to linux-stable.git: 26 size=["unknown", "unknown"]
requests to linux.git: 38 size=["unknown", "unknown"]
requests to chromiumos: 59 size=["unknown", "unknown"]
requests to lineageOS: 59 size=["unknown", "unknown"]
requests to linux-next.git: 302 size=["unknown", "unknown"]
requests to chakra-releases: 1774 size=["68.99G", "68.99G"]
requests to bananian: 1887 size=["unknown", "2.78G"]
requests to dotdeb: 2146 size=["25.22G", "25.22G"]
requests to slackwarearm: 2233 size=["28.84G", "28.84G"]
requests to bjlx: 2486 size=["3.63G", "3.63G"]
requests to pybombs: 2723 size=["unknown", "unknown"]
requests to openresty: 2899 size=["unknown", "5.17G"]
requests to elvish: 3050 size=["1.65G", "1.65G"]
requests to loongson: 3796 size=["16.87G", "16.87G"]
requests to gnu-alpha: 4351 size=["30.63G", "30.63G"]
requests to dell: 5362 size=["933.77G", "874G"]
requests to AOSP: 6266 size=["unknown", "unknown"]
requests to gentoo-portage-prefix: 7911 size=["206.43M", "206.43M"]
requests to kali-security: 9242 size=["128K"]

[jiegec]

Propose 一个删除的流程：

1. 选择候选的 repo ：最近三个月访问量小于一个阈值
2. 在本仓库新建一个相应的 issues，开始投票。
3. 指定的 TUNA 成员拥有一票否决权，通过关闭 issue 行使；其他成员享有正常的投票权，通过 Reaction 来投票
4. 一段固定的时间后，统计投票结果，如果同意删除镜像占总投票人数过半，则删除该镜像

[z4yx]

上面的统计可能要去掉一些刚加入的镜像。

现在 dell 的上游已经无法连接，如果没有用户，可以考虑删除了。

[jiegec]

但是 dell 上游的 http 还是在的

[abcfy2]

排第一的是crates.io-index.git，这个镜像我之前还用过，不知道为什么tuna的特别慢，经常错误重试，于是我就改到了ustc的git协议的镜像上去了。

不过话说回来，rust在国内确实冷门，所以crates.io-index镜像使用率确实低

[z4yx]

前面列出的镜像又部分是最近刚加上了，所以在三个月的统计排名中显得很低。

---

已更新列表。

[z4yx]

关于Git出错的问题，之前也有很多人报告过。我们从校内无法复现这个问题，不知道是哪一环的网络造成了这个故障。

[ttys3]

crates.io-index.git 希望能够保留

[z4yx]

之前的数据已经过时了，@jiegec 有空时重新统计下吧

[jiegec]

elasticsearch: showing 2020.0*.*
requests to glibc.git: 8 size=["193M", "193M"]
requests to openthos-src: 8 size=["27G", "27G"]
requests to binutils-gdb.git: 12 size=["384M", "384M"]
requests to linux-firmware.git: 12 size=["313M", "313M"]
requests to linux-next.git: 12 size=["1.7G", "1.7G"]
requests to linux-stable.git: 14 size=["2.1G", "2.1G"]
requests to flutter-sdk.git: 16 size=["371M", "371M"]
requests to gentoo-portage.git: 18 size=["3.5G", "3.5G"]
requests to qemu.git: 20 size=["190M", "190M"]
requests to gcc.git: 34 size=["1.6G", "1.6G"]
requests to crates.io-index.git: 36 size=["346M", "346M"]
requests to lineageOS: 50 size=["236G", "236G"]
requests to linux.git: 55 size=["1.7G", "1.7G"]
requests to AOSP: 92 size=["371G", "371G"]
requests to llvm: 92 size=["943M", "943M"]
requests to CocoaPods: 105 size=["820M", "820M"]
requests to homebrew: 1120 size=["1.5G", "1.5G"]
requests to dotdeb: 1868 size=["25.22G", "23.8G"]
requests to pybombs: 2612 size=["2.3G", "2.3G"]
requests to elvish: 2903 size=["1.83G", "1.72G"]
requests to bananian: 3320 size=["unknown", "2.78G"]
requests to bjlx: 3930 size=["3.84G", "3.84G"]
requests to openresty: 5597 size=["unknown", "8.78G"]
requests to nix: 5863 size=["unknown", "1.72G"]
requests to gnu-alpha: 6927 size=["30.96G", "30.96G"]
requests to chakra-releases: 7117 size=["68.99G", "68.99G"]
requests to loongson: 7165 size=["16.39G", "16.39G"]
requests to HHVM: 8970 size=["1.28T", "1.17T"]

[z4yx]

对git仓库请求数的统计是不是有错误，不可能只有这么几次的。

[jiegec]

应该是没收集到 git 的 log 吧