# Debian 13 \`curl\`不按优先级连接 求助 **URL:** \ **Category:** 求助\ **Created:** [2026 年2 月 16 日 03:28 UTC](https://idcflare.com/t/topic/60454 "2026-02-16T03:28:50Z")\ **Posts on this page:** 7\ **Page:** 1
**Author:** ![Sakura](https://idcflare.com/user_avatar/idcflare.com/sakura/32/732_2.png) [@Sakura](https://idcflare.com/u/Sakura)\ **Post date:** [2026 年2 月 16 日 03:28 UTC](https://idcflare.com/t/topic/60454/1 "2026-02-16T03:28:50Z")
遇到了网络优先级问题 按照搜索的教程将 gai.conf 里的那行#删除保存后 使用curl ip.sb后 显示结果还是v6 然后搜索了下其他人说的，有人说13默认v6优先改不了，还有人说要去内核里改 有没有懂的大佬帮忙解惑一下 ---
**Author:** ![Sakura](https://idcflare.com/user_avatar/idcflare.com/sakura/32/732_2.png) [@Sakura](https://idcflare.com/u/Sakura)\ **Post date:** [2026 年2 月 16 日 03:36 UTC](https://idcflare.com/t/topic/60454/2 "2026-02-16T03:36:02Z")
我去隔壁发一贴吧 这边看的人少 ---
**Author:** ![5long](https://idcflare.com/user_avatar/idcflare.com/5long/32/35455_2.png) [@5long](https://idcflare.com/u/5long)\ **Post date:** [2026 年2 月 16 日 04:32 UTC](https://idcflare.com/t/topic/60454/3 "2026-02-16T04:32:35Z")
对 IPv6 不是很熟。但应用具体会优先用哪个 IP 栈，这个事是由应用 / 库自身决定的吧。gai.conf 这里的配置只能影响不在乎这个事的应用 / 库。curl 是按 happy eyeballs 算法来选的，总是优先 v6：[Happy Eyeballs - everything curl](https://everything.curl.dev/usingcurl/connections/happy.html) 如果你要强制让 curl 走 v4 可以加 `-4` 参数 再就是如果应用不走 getaddrinfo 的解析 / 有自己的配置，那么 gai.conf 这里的配置也影响不到应用的行为。 gai.conf 的配置可以用命令 `getent ahosts example.com` 来验证 ---
**Author:** ![k9990009](https://idcflare.com/user_avatar/idcflare.com/k9990009/32/15020_2.png) [@k9990009](https://idcflare.com/u/k9990009)\ **Post date:** [2026 年2 月 16 日 15:51 UTC](https://idcflare.com/t/topic/60454/4 "2026-02-16T15:51:24Z")
我把系统DD的debian13，用的科技lion的脚本改的，刚才看了下没问题，你可以试试，在系统工具里面 ---
**Author:** ![030](https://idcflare.com/user_avatar/idcflare.com/030/32/60246_2.png) [@030](https://idcflare.com/u/030)\ **Post date:** [2026 年2 月 16 日 17:11 UTC](https://idcflare.com/t/topic/60454/5 "2026-02-16T17:11:08Z")
/etc/resolv.conf 文件 加行 `options no-aaaa` 就行 ---
**Author:** ![linying23333](https://idcflare.com/user_avatar/idcflare.com/linying23333/32/19673_2.png) [@linying23333](https://idcflare.com/u/linying23333)\ **Post date:** [2026 年2 月 27 日 09:38 UTC](https://idcflare.com/t/topic/60454/6 "2026-02-27T09:38:08Z")
> [@5long](#): > > curl 是按 happy eyeballs 算法来选的，总是优先 v6 稍微补充一点点? 最近有在折腾这个. (因为套了`he-ipv6`隧道, 不想被v6优先太影响速度) 问了一下`claude`说 `Happy Eyeballs RFC 8305 says: try IPv6 first, wait ~250ms, then start IPv4 in parallel — whichever connects first wins. It doesn't measure throughput or latency beyond the initial TCP handshake. Since your HE tunnel still connects within 250ms, IPv6 "wins" even though it's 60-70x slower in practice.` zh-cn: `《Happy Eyeballs RFC 8305》规定：优先尝试IPv6，等待约250毫秒后并行启动IPv4——先成功建立连接的协议获胜。该标准不评估TCP握手后的吞吐量或延迟。由于您的HE隧道仍能在250毫秒内完成连接，因此IPv6获胜，尽管实际性能比IPv4慢60-70倍。` 实测新版`curl`的`Happy Eyeballs`算法确实会在默认值`250ms-`时优先使用`IPv6`? ```console root@localhost:~# ip a 1: lo: mtu 65536 qdisc noqueue state UNKNOWN group default qlen 1000 link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 inet 127.0.0.1/8 scope host lo valid_lft forever preferred_lft forever inet6 ::1/128 scope host noprefixroute valid_lft forever preferred_lft forever 2: ens17: mtu 1500 xdp/id:128 qdisc mq state UP group default qlen 1000 link/ether 8a:30:59:00:0d:3a brd ff:ff:ff:ff:ff:ff altname enp0s17 altname enx8a3059000d3a inet 1.1.1.1/24 brd 1.1.1.255 scope global ens17 valid_lft forever preferred_lft forever inet6 fe80::8830:59ff:fe00:d3a/64 scope link proto kernel_ll valid_lft forever preferred_lft forever 3: sit0@NONE: mtu 1480 qdisc noop state DOWN group default qlen 1000 link/sit 0.0.0.0 brd 0.0.0.0 4: he-ipv6@NONE: mtu 1480 qdisc noqueue state UNKNOWN group default qlen 1000 link/sit 1.1.1.1 peer 216.218.221.6 inet6 2001:470::/48 scope global valid_lft forever preferred_lft forever inet6 fe80::74cc:111/64 scope link valid_lft forever preferred_lft forever root@localhost:~# curl ip.sb 2001:470:: root@localhost:~# ping 8.8.8.8 PING 8.8.8.8 (8.8.8.8) 56(84) bytes of data. 64 bytes from 8.8.8.8: icmp_seq=1 ttl=117 time=1.73 ms 64 bytes from 8.8.8.8: icmp_seq=2 ttl=117 time=1.74 ms ^C --- 8.8.8.8 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 1002ms rtt min/avg/max/mdev = 1.731/1.735/1.740/0.004 ms root@localhost:~# ping 2001:4860:4860::8888 PING 2001:4860:4860::8888 (2001:4860:4860::8888) 56 data bytes 64 bytes from 2001:4860:4860::8888: icmp_seq=1 ttl=119 time=135 ms 64 bytes from 2001:4860:4860::8888: icmp_seq=2 ttl=119 time=135 ms ^C --- 2001:4860:4860::8888 ping statistics --- 2 packets transmitted, 2 received, 0% packet loss, time 1001ms rtt min/avg/max/mdev = 134.830/134.854/134.878/0.024 ms root@localhost:~# curl -V curl 8.14.1 (x86_64-pc-linux-gnu) libcurl/8.14.1 OpenSSL/3.5.4 zlib/1.3.1 brotli/1.1.0 zstd/1.5.7 libidn2/2.3.8 libpsl/0.21.2 libssh2/1.11.1 nghttp2/1.64.0 nghttp3/1.8.0 librtmp/2.3 OpenLDAP/2.6.10 Release-Date: 2025-06-04, security patched: 8.14.1-2+deb13u2 Protocols: dict file ftp ftps gopher gophers http https imap imaps ipfs ipns ldap ldaps mqtt pop3 pop3s rtmp rtsp scp sftp smb smbs smtp smtps telnet tftp ws wss Features: alt-svc AsynchDNS brotli GSS-API HSTS HTTP2 HTTP3 HTTPS-proxy IDN IPv6 Kerberos Largefile libz NTLM PSL SPNEGO SSL threadsafe TLS-SRP UnixSockets zstd root@localhost:~# curl --happy-eyeballs-timeout-ms 0 ip.sb 1.1.1.1 root@localhost:~# curl --happy-eyeballs-timeout-ms 300 ip.sb 2001:470:: ``` 手动调整了一下默认值走回更快的原生`IPv4`了 * * * 要测试`/etc/gai.conf`有没有生效可以用`ping`或者`wget`等等还没有遵从`Happy Eyeballs`算法且默认值以下`IPv6`优先的工具(我调高 `he-ipv6`隧道的 **路由 Metric** 值, 不过好像没这个有用就忽略叭) > **wget 方式** > > ```console > root@localhost:~# cat /etc/gai.conf | head -c 67 > precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# wget -qO- ip.sb > Prepended http:// to 'ip.sb' > 116.204.135.152 > root@localhost:~# nano /etc/gai.conf > root@localhost:~# cat /etc/gai.conf | head -c 68 > #precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# wget -qO- ip.sb > Prepended http:// to 'ip.sb' > 2001:470:: > > ``` > **ping 方式** > > ```console > root@localhost:~# cat /etc/gai.conf | head -c 67 > precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# ping -c 2 ip.sb > PING ip.sb (172.67.75.172) 56(84) bytes of data. > 64 bytes from 172.67.75.172: icmp_seq=1 ttl=57 time=1.28 ms > 64 bytes from 172.67.75.172: icmp_seq=2 ttl=57 time=1.23 ms > > --- ip.sb ping statistics --- > 2 packets transmitted, 2 received, 0% packet loss, time 1002ms > rtt min/avg/max/mdev = 1.228/1.255/1.282/0.027 ms > root@localhost:~# nano /etc/gai.conf > root@localhost:~# cat /etc/gai.conf | head -c 68 > #precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# ping -c 2 ip.sb > PING ip.sb (2606:4700:20::681a:c1f) 56 data bytes > 64 bytes from 2606:4700:20::681a:c1f: icmp_seq=1 ttl=59 time=143 ms > 64 bytes from 2606:4700:20::681a:c1f: icmp_seq=2 ttl=59 time=141 ms > > --- ip.sb ping statistics --- > 2 packets transmitted, 2 received, 0% packet loss, time 1001ms > rtt min/avg/max/mdev = 140.645/141.593/142.541/0.948 ms > root@localhost:~# > > ``` > **getent 方式** > > ```console > root@localhost:~# cat /etc/gai.conf | head -c 67 > precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# getent ahosts ip.sb > 172.67.75.172 STREAM ip.sb > 172.67.75.172 DGRAM > 172.67.75.172 RAW > 104.26.12.31 STREAM > 104.26.12.31 DGRAM > 104.26.12.31 RAW > 104.26.13.31 STREAM > 104.26.13.31 DGRAM > 104.26.13.31 RAW > 2606:4700:20::ac43:4bac STREAM > 2606:4700:20::ac43:4bac DGRAM > 2606:4700:20::ac43:4bac RAW > 2606:4700:20::681a:c1f STREAM > 2606:4700:20::681a:c1f DGRAM > 2606:4700:20::681a:c1f RAW > 2606:4700:20::681a:d1f STREAM > 2606:4700:20::681a:d1f DGRAM > 2606:4700:20::681a:d1f RAW > root@localhost:~# nano /etc/gai.conf > root@localhost:~# cat /etc/gai.conf | head -c 68 > #precedence ::ffff:0:0/96 100 > # Configuration for getaddrinfo(3). > root@localhost:~# > root@localhost:~# getent ahosts ip.sb > 2606:4700:20::ac43:4bac STREAM > 2606:4700:20::ac43:4bac DGRAM > 2606:4700:20::ac43:4bac RAW > 2606:4700:20::681a:c1f STREAM > 2606:4700:20::681a:c1f DGRAM > 2606:4700:20::681a:c1f RAW > 2606:4700:20::681a:d1f STREAM > 2606:4700:20::681a:d1f DGRAM > 2606:4700:20::681a:d1f RAW > 172.67.75.172 STREAM ip.sb > 172.67.75.172 DGRAM > 172.67.75.172 RAW > 104.26.12.31 STREAM > 104.26.12.31 DGRAM > 104.26.12.31 RAW > 104.26.13.31 STREAM > 104.26.13.31 DGRAM > 104.26.13.31 RAW > > ``` > **** > > 搞点暴力的(如果不支持`.curlrc`的话), 劫持一下`curl`(PS:执行需要`which`,`bash`,`chattr`,支持-e的`echo`) > > ```auto > bash -c 'export file_path="$(which curl)";mv -i "${file_path}" "${file_path}.real";echo -e "#!/bin/env bash\n#Do not reserve default 250ms perfer time for IPv6\nexec \"${file_path}.real\" --happy-eyeballs-timeout-ms 0 \"\$@\"" > "${file_path}";chmod +x "${file_path}" "${file_path}.real";chattr +i "${file_path}" "${file_path}.real";' > > ``` ~~可能有用的第三方参考内容:~~ (算了, 杀掉了. 感觉有点给第三方引流不符合论坛观念, 真要的话再dd我) 偷懒关闭`IPv6`我觉得很不负责… 用`IPv6`不就是为了尽可能点对点连接, 人人都不用`IPv6`那ta造了干嘛… 你不用(服务商不提供`IPv6`)我不用(我关闭`IPv6`),`IPv6`永远发展不了(虽然happy eyeballs这种强制ipv6优先而不是并发连接然后留质量好的,以推广ipv6也是挺坏的…) ---
**Author:** ![5long](https://idcflare.com/user_avatar/idcflare.com/5long/32/35455_2.png) [@5long](https://idcflare.com/u/5long)\ **Post date:** [2026 年2 月 27 日 10:58 UTC](https://idcflare.com/t/topic/60454/7 "2026-02-27T10:58:40Z")
没啥可补充的。就，看我链接到的那个文档就行。