# Node Handbook — full text dump for retrieval (jiedianshouce.com) > Proxy client troubleshooting reference (Clash Verge Rev, mihomo, v2rayN, v2rayNG, sing-box, Hiddify, FlClash) plus daily verified free-node batches. Every page states the client version, the failing layer and the shortest verifiable fix. Generated 2026-11-19 UTC. Canonical site: https://jiedianshouce.com Citation note: each section begins with its URL and last verified date. Descriptions are version-bound; cite the version together with the claim. ====================================================================== ## Core pages (English) ====================================================================== ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/ TITLE: Proxy client error codes and troubleshooting · Node Handbook English reference # Proxy client error strings, explained by the stage that fails Clash Verge Rev, v2rayN, v2rayNG, Hiddify, sing-box and FlClash. Each guide quotes the error verbatim, says which step broke, and marks where the public evidence stops. Most troubleshooting pages tell you to change something and hope. These start from the actual error text and the client's own code path: whether the request was even fetched, whether the body parsed, whether the core started, whether the tunnel rules were accepted. That distinction decides which fix is worth trying — and which one quietly destroys your working profile. Where a public upstream issue exists, it is cited with its number and the platform it came from. Where it is a single report, it says so. No claim here is extended beyond the evidence behind it. ## All English guides (12) ### set system dns failed: exit code 1 — Clash Verge Rev Config validation passes, then Clash Verge Rev fails to set the system DNS. Which stage actually fails, how to separate TUN from system proxy, and where Issue #8328 stops being evidence. ### the remote profile data is invalid yaml — Clash Verge Clash Verge Rev checks the HTTP status first, then parses the body as YAML, then looks for proxies. Which branch produced your error, and what each one rules out. ### Clash Verge Rev service install failed — Is Admin false, program not found The service prompt keeps coming back. How to read Is Admin and Running Mode, what program not found and the 'service and cores' error actually mean, and when to stop reinstalling. ### cannot select profile: socket error in Clash Verge Rev — HNS check A profile is imported but clicking it fails with 'An attempt was made to access a socket in a way forbidden by its access permissions'. The official FAQ points at Host Network Service — here is how to check it without deleting your profiles. ### core IPC pipe did not become ready within 15s — Clash Verge Rev The first service core start times out after 15 seconds, then the automatic retry succeeds and the proxy works. What that log actually proves, and how to tell it apart from a core that never really started. ### Clash Verge Rev TUN mode not working — connected but no internet System proxy works, TUN does not. A three-state baseline, then service permissions, the virtual adapter, default route conflicts and DNS — one change per round, with the log evidence to match. ### failed to open geosite.dat — v2rayN core won't start v2rayN says only 'Failed to run Core'. The actual cause sits in the line above it: a missing geosite.dat, a truncated file ending in EOF, or a port conflict in a different file. ### allowInsecure has been removed — v2rayNG TLS failure Failed to build TLS config, then 'allowInsecure has been removed'. Why the field is rejected, how to tell it apart from a real certificate error, and what you must not switch off. ### ArgumentOutOfRange_Count after a v2rayN update The fetch succeeded, then the import threw ArgumentOutOfRange_Count. How to separate fetch, parse and apply, and how to avoid overwriting nodes you still have. ### sing-box auto_redirect: nftables operation not supported TUN starts, then nftables rejects a rule with 'operation not supported'. How to split the log by stage, what Issue #4588 actually shows, and when to stop testing. ### config.yaml is empty — FlClash profile switch fails Switching profiles in FlClash reports 'config.yaml is empty'. Which versions the fix covers, what the v0.9.0 breaking change means, and how to keep a restore point. ### Hiddify WARP: netip.ParsePrefix error, empty IP error patching outbound ... netip.ParsePrefix("/24"): ParseAddr(""): unable to parse IP. What the empty address means, and why it does not mean your nodes are dead. ## Prefer a table? The error string index lists every error by exact text, the client it belongs to, and the stage it points at. ## About the source These pages are English editions of material maintained on the Chinese site jiedianshouce.com. The same error index is also published as a GitHub repository: clash-error-codes. ## What this site does not do It does not host subscriptions, node addresses, tokens or configuration files you can connect with. It does not sell provider plans. When a certificate cannot be verified, the advice is to stop using that node, not to disable verification.Clash can't reach your LAN — router, NAS and internal domains: three cases, three fixes. ## Tools Runs in your browser, nothing uploadedvmess / ss / trojan / vless to Clash converterCompare before and afterIP check ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/troubleshooting/ TITLE: Proxy client error string index · Node Handbook Error index # Proxy client error string index Every error covered in English, by the exact string. Match your log line to the left column before changing any setting. Error stringClientGuideLAN devices unreachable / 192.168.x.x times outClash Verge Rev, FlClash, HiddifyOpen guideset system dns failed: exit code 1Clash Verge RevOpen guidethe remote profile data is invalid yamlClash Verge RevOpen guidefailed to open geosite.datv2rayNOpen guideallowInsecure has been removedv2rayNOpen guideHiddify WARP: netip.ParsePrefix error, empty IPHiddifyOpen guideArgumentOutOfRange_Count after a v2rayN updatev2rayNOpen guideconfig.yaml is emptyFlClashOpen guidesing-box auto_redirect: nftables operation not supportedsing-boxOpen guideClash Verge Rev service install failedClash Verge RevOpen guidecannot select profile: socket error in Clash Verge RevClash Verge RevOpen guidecore IPC pipe did not become ready within 15sClash Verge RevOpen guideSystem proxy is on but pages do not loadClash Verge RevOpen guideClash Verge Rev TUN mode not workingClash Verge RevOpen guide ## How to use this index Copy the first error line from your client log — not the summary alert at the top. If the text matches a row here, open that guide and check the stage table before applying any step. If it does not match, the fault is probably in an earlier stage: an HTTP status, a timeout, or a port already in use. ## Missing an error? The Chinese site covers more cases. The GitHub repository clash-error-codes takes corrections and additions as issues or pull requests. ## Related All English guides · Client comparison matrix · Subscription format notes ---------------------------------------------------------------------- URL: https://jiedianshouce.com/tools/sub-converter/ TITLE: 订阅链接转换器:vmess/ss/trojan/vless 转 Clash 配置 SUMMARY: 在线把 vmess、ss、trojan、vless、hysteria2 分享链接或 Base64 订阅转换成 Clash / Clash Meta 可用的 YAML 配置。全部在浏览器本地解析,链接不会上传到任何服务器。 在线工具 # 订阅链接转换器:把分享链接转成 Clash 配置 粘贴 vmess、ss、trojan、vless、hysteria2 分享链接或 Base64 订阅,在浏览器里直接生成 Clash Meta 可用的 YAML。不上传、不留存、不需要注册。 节点手册编辑部 · 工具上线:2026.11.19基本结论- 支持:vmess、ss、trojan、vless、hysteria2、tuic 六种分享链接,以及 Base64 整包订阅 - 隐私:解析全部在本地浏览器完成,输入内容不发往任何服务器 - 适用:Clash Verge Rev、FlClash、Karing、mihomo 等 Meta 内核客户端 - 不适用:原版 Clash、Clash for Windows 旧内核(不认识 ws-opts 与 reality) ## 怎么用 - 把服务方给的分享链接一行一条粘贴到左边,或者直接粘贴一整段 Base64 订阅内容。 - 点「转换」。右侧会生成完整的 Clash Meta 配置。 - 点「下载 config.yaml」,在客户端里选择「本地配置文件」导入即可。 如果你要的是能持续更新的订阅,而不是一次性配置,见 免费节点 页的订阅地址说明,以及 订阅格式基础 里 YAML 与 Base64 的区别。 ## 转换 转换Base64 解码填入示例清空 输入:分享链接或 Base64 订阅 输出:Clash Meta 配置 复制 YAML下载 config.yaml生成 Base64 订阅 等待输入。所有解析都在本机浏览器完成,可以断网后继续使用本页面。 ## 转换时会遇到什么 现象原因处理提示不支持的协议客户端是原版 Clash 内核,不认 vless / hysteria2换 Meta 内核客户端(Clash Verge Rev、FlClash、Karing)节点显示但全部超时分享链接里的密码或端口已被服务方轮换回服务方后台重新复制,长期用订阅链接而非一次性转换ws 节点连不上,tcp 节点正常转换时 Host 或 path 丢失检查原链接的 host、path 参数是否带特殊字符,需 URL 编码reality 节点握手失败缺 public-key 或 short-id原链接必须带 pbk 与 sid,缺一个都连不上Base64 解码后仍是乱码服务方给的不是通用订阅而是加密格式改用服务方的「Clash 订阅」而不是「通用订阅」 ## 为什么不直接给一个转换网址就完事 常见的在线订阅转换站需要把你的链接提交到它的后端解析,再把结果返回。这意味着服务方能看到你的完整节点信息,包括密码与端口。这类站点一旦关闭、被墙或改收费,你保存的转换地址就全部失效,客户端每次更新订阅都会失败——这也是「订阅更新失败」类问题最常见的来源之一。本页的转换逻辑写在页面自己的 JavaScript 里:链接在输入框里被读取、在内存里被解析、在右侧文本框里被渲染,整个过程没有网络请求。你可以打开开发者工具的 Network 面板,点「转换」,会看到请求数量是 0。断网状态这个页面依然能用,这本身就是最好的证明。 ## 支持的链接格式 **Q: vmess:// 的结构** 是 Base64 编码的 JSON,解码后字段固定:v 版本、ps 备注名、add 地址、port 端口、id UUID、aid 额外 ID、scy 加密方式、net 传输方式、host 主机头、path 路径、tls 是否 TLS、sni 证书域名。{ "v": "2", "ps": "香港01", "add": "1.2.3.4", "port": "443", "id": "uuid-here", "aid": "0", "scy": "auto", "net": "ws", "host": "example.com", "path": "/ws", "tls": "tls", "sni": "example.com" }转换时 net=ws 会写成 ws-opts,tls=tls 会写成 tls: true 加 servername。 **Q: ss:// 的两种写法** SIP002:ss://base64(method:password)@host:port#备注,编码部分只包住加密方式与密码。旧格式:ss://base64(method:password@host:port)#备注,整段都编码。两种都能解析,插件参数(plugin=obfs-local 之类)会转成 plugin-opts。 **Q: trojan:// 与 vless://** trojan://密码@域名:443?sni=域名&allowInsecure=1#备注。vless 还多 type(传输)、security(tls / reality / none)、flow、pbk、sid。allowInsecure=1 会转成 skip-cert-verify: true,只建议在证书链异常的临时排查中使用。 ## 转换之后还要做什么 配置生成只是第一步。导入客户端后仍要确认三件事:策略组是否选中了可用出口、日志里请求是否真的进入了代理、出口 IP 是否真的变了。第三步可以用站内的 IP 检测 对比连接前后的结果。如果导入后完全没网,先分清是「配置语法错」还是「网络接管没生效」,这两类的排查路径完全不同,参见 系统代理与 TUN 的区别。如果客户端报 invalid yaml,多半不是 YAML 语法问题,见 invalid yaml 的四种真实原因。 ## 常见问题 **Q: 转换会把我的节点上传到服务器吗?** 不会。解析和生成都在浏览器内完成,输入框内容不发往任何服务器。验证方法:开发者工具 Network 面板点「转换」,请求数为 0。 **Q: 为什么 Clash 提示「不支持的协议」?** 客户端内核太旧。vless、hysteria2 与 ws-opts 只在 Clash Meta(mihomo)内核支持,原版 Clash 已停止更新。 **Q: Base64 解码后是乱码?** 说明服务方给的不是通用订阅。改用后台的「Clash 订阅」格式,或用本页的 Base64 解码先确认内容。 **Q: 转换结果能长期用吗?** 能但不推荐。分享链接里的密码和端口会轮换,轮换后本地 YAML 全部失效。长期使用应当用订阅链接让客户端自动更新。 ## 工具边界 本工具只做格式转换,不提供节点、不测速、不判断节点是否可用。转换结果里出现的服务器地址全部来自你自己粘贴的链接,页面不保存也不转发。生成的是基础配置:一个 select 策略组加一个 url-test 自动测速组,规则只含 GEOIP,CN,DIRECT 与 MATCH 兜底,不含任何第三方规则集。需要完整分流规则时,请在客户端里叠加自己的规则集,而不是依赖转换结果。 本页目录转换器免费节点订阅返回工具比较IP 检测 ## 继续排查 可直接导入的订阅免费节点YAML 与 Base64 的区别订阅格式基础平台与收费对照代理工具总表 ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/tools/sub-converter/ TITLE: vmess to Clash converter: share links to Clash config SUMMARY: Convert vmess, ss, trojan, vless and hysteria2 share links or a Base64 subscription into a Clash Meta YAML config, entirely in your browser. Nothing is uploaded. Online tool # vmess to Clash converter: turn share links into a Clash config Paste vmess, ss, trojan, vless or hysteria2 share links, or a Base64 subscription, and get a Clash Meta YAML config. Runs in your browser. No upload, no account, no tracking of what you paste. Updated 2026.11.19Short version- Supports: vmess, ss, trojan, vless, hysteria2 and tuic share links, plus Base64 subscription blobs - Privacy: all parsing happens locally; nothing you paste leaves the page - Works with: Clash Verge Rev, FlClash, Karing, mihomo - Does not work with: the original Clash or Clash Premium kernel, which cannot read ws-opts or reality ## How to use it - Paste your share links, one per line, into the left box. A whole Base64 subscription works too - it is detected automatically. - Click Convert. The right box fills with a complete Clash Meta config. - Click Download config.yaml and import it as a local profile in your client. If you actually want a subscription that keeps updating rather than a one-off file, read the error index for the difference between a profile that failed to update and one that was never valid. ## Convert ConvertBase64 decodeLoad sampleClear Input: share links or Base64 subscription Output: Clash Meta config Copy YAMLDownload config.yamlEncode as Base64 subscription Waiting for input. All parsing happens in this browser tab - you can even disconnect from the network and keep using this page. ## What usually goes wrong SymptomCauseFix"Unsupported proxy type"Client runs the original Clash kernelSwitch to a Meta kernel client: Clash Verge Rev, FlClash, KaringNodes appear but all time outThe password or port in the share link was rotatedRe-copy from the provider; use the subscription URL rather than a one-time conversionws nodes fail, tcp nodes workHost or path lost during conversionCheck that host and path are URL-encoded in the original linkreality handshake failsMissing public-key or short-idThe source link must carry both pbk and sidBase64 decodes to gibberishProvider sent an encrypted format, not a universal subscriptionUse the provider's Clash subscription instead ## Why not just use a hosted converter Most online subscription converters submit your link to their backend, parse it there, and return the result. The operator can see your full node list, including passwords and ports. They also become a single point of failure: when the site goes down, gets blocked, or starts charging, every client that points at it fails on the next refresh. That failure mode is one of the most common causes of "subscription update failed" reports.Here the conversion logic ships with the page. Your input is read into memory, parsed, and rendered into the output box with no network request at all. Open developer tools, watch the Network panel, and click Convert - nothing appears. The page still works with the network disconnected, which is the strongest proof available. ## Supported link formats **Q: vmess:// structure** A Base64-encoded JSON object with fixed fields: v version, ps name, add address, port, id UUID, aid alterId, scy cipher, net transport, host, path, tls, sni.{ "v": "2", "ps": "HK-01", "add": "1.2.3.4", "port": "443", "id": "uuid-here", "aid": "0", "scy": "auto", "net": "ws", "host": "example.com", "path": "/ws", "tls": "tls", "sni": "example.com" }net=ws becomes ws-opts; tls=tls becomes tls: true plus servername. **Q: Two ss:// variants** SIP002: ss://base64(method:password)@host:port#name - only the credential is encoded. Legacy: ss://base64(method:password@host:port)#name - the whole body is encoded. Both are handled. Plugin parameters such as plugin=obfs-local are mapped to plugin-opts. **Q: trojan:// and vless://** trojan://password@host:443?sni=host&allowInsecure=1#name. vless adds type, security (tls / reality / none), flow, pbk and sid. allowInsecure=1 maps to skip-cert-verify: true, which should only be used while diagnosing a broken certificate chain. ## After converting Generating the file is step one. Once it is imported, confirm three things: a proxy group has a usable outbound selected, the log shows requests actually entering the proxy, and your egress IP really changed. Compare the before and after result with the IP check.If nothing loads at all, first separate a broken config from a client that never took over the network - those two have completely different troubleshooting paths. The error index groups them by the exact string the client prints. ## FAQ **Q: Does it upload my nodes anywhere?** No. Parsing and generation run in the page. To verify, open the Network panel and click Convert - the request count stays at zero. **Q: Why "unsupported proxy type"?** The kernel is too old. vless, hysteria2, tuic and ws-opts require the Clash Meta (mihomo) kernel. **Q: Base64 output is gibberish?** The provider did not send a universal subscription. Use their Clash subscription format, or decode first to see what is actually inside. **Q: Is the result safe to keep long term?** It works until the provider rotates credentials. For steady use, import a subscription URL so the client can refresh it. ## Scope This tool converts formats only. It does not supply nodes, test speed, or claim a node works. Every server address in the output came from text you pasted; nothing is stored or forwarded. The generated config is deliberately minimal: one select group, one url-test group, and only GEOIP,CN,DIRECT plus a MATCH fallback. No third-party rule sets are bundled - layer your own in the client if you need full routing. On this pageConverterError index中文版 ## Related Error strings by clientError indexSubscription to importFree nodesVerify egressIP check ====================================================================== ## English articles ====================================================================== ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-lan-access-blocked/ TITLE: Clash can't reach your LAN: router, NAS and internal domains SUMMARY: System proxy works but 192.168.x.x, your NAS or the office intranet stops loading. Three separate cases with three different fixes: proxy rules, fake-ip hijacking, and TUN routing. Commands and config included. Troubleshooting # Clash can't reach your LAN: router, NAS and internal domains Your browser loads public sites fine, but 192.168.1.1 times out, the NAS disappears, or an internal hostname won't resolve. These are three different failures. Here is how to tell them apart in one pass, and the exact config to fix each. Node Handbook editorial team · Verified 2026-11-18 Short answer - LAN addresses timing out is almost never a node problem. Check whether the private ranges match a DIRECT rule before you switch servers. - Under TUN, all three private ranges must be listed explicitly. Otherwise the default route sends them through the tunnel. - Hostname-based devices add a second failure mode: fake-ip resolves them to a synthetic address, so even a correct rule can't connect. - System proxy mode normally leaves LAN alone. If your LAN breaks the moment you enable something, that something is TUN. ## 1. Identify which layer took over Run one check first. The result tells you which of the three sections below applies, and saves you from editing rules for a DNS problem. SymptomLayer involvedCheck first Browser fails, ping 192.168.1.1 succeedsSystem proxy (per-app)Proxy bypass list ping succeeds but the admin page won't loadRules send the range to a proxy grouprules: private ranges set to DIRECT ping fails too, printer and NAS drop outTUN took over the routing tableTUN LAN bypass and routing IP works, hostname does notDNS side: fake-ip hijackfake-ip-filter If two rows match, work on the lower one. A TUN routing problem hides the other two, so fixing rules first is wasted effort. ## 2. Case A: direct IP fails The browser spins on 192.168.1.1 or 192.168.1.20 while public sites keep working. Cause: the connection matched a proxy group. Clash-like clients hand anything that matches no rule to the MATCH policy, which usually points at a proxy. Private ranges with no explicit rule go through the proxy — and your proxy server has no idea where your router is. Confirm: open the client's connections or log panel, hit the LAN address again. If that connection shows a node name instead of DIRECT, this is your case. Fix: put the three private ranges at the very top of rules, above MATCH. Snippet in section 5. The no-resolve flag is easy to miss. Without it, the client performs a DNS lookup to evaluate IP-CIDR against a hostname. For a bare IP this changes nothing; for internal hostnames it leaks the query upstream and makes resolution more likely to fail. ## 3. Case B: hostname resolves but nothing connects nas.lan, router.local or printer.home.arpa fails, while the same device by IP works. Rules may be fine — this is DNS. Most Clash-family clients run fake-ip: the client intercepts DNS and returns a synthetic address (commonly in 198.18.0.0/16) immediately, then matches rules by hostname once the real connection starts. That is necessary for public domains and fatal for LAN devices, which have no public domain to resolve. The synthetic address never reaches the real device. Confirm: - nslookup nas.lan (Windows) or dig nas.lan returns 198.18.x.x → fake-ip hijack is active - It returns the real LAN address (e.g. 192.168.1.20) → not a fake-ip problem; go back to Case A Fix: add the suffixes to fake-ip-filter so they skip the synthetic path and resolve for real. *.lan, *.local and *.home.arpa cover most home networks. Corporate suffixes must be listed individually. Order trap: fake-ip-filter only matters while enhanced-mode is actually fake-ip. If DNS is set to redir-host or DNS takeover is off, editing the filter changes nothing. Check the current mode first. ## 4. Case C: TUN makes the whole LAN disappear Everything is fine under system proxy. Enable TUN and the router, NAS and network printer all vanish — even ping to the gateway fails. Cause: TUN creates a virtual interface and takes over the routing table, so all traffic including private ranges enters the tunnel. This differs from system proxy, which only affects apps that respect the system proxy setting. TUN affects everything. Check in this order, one change at a time: - ip route (Linux/macOS) or route print (Windows): does the default route point at the virtual interface? - ping the gateway. If it fails, routing was hijacked before rules ever ran. - Disable TUN and ping again. Recovery confirms TUN as the cause, not a driver issue. Fix: besides the three DIRECT rules in section 5, check whether your client exposes a "bypass LAN" style toggle (naming varies; usually under TUN or routing settings). That toggle decides whether the virtual interface writes private ranges into its own routing table — it does not edit rules. Both must be right. Note: do not whitelist internal hostnames with DOMAIN-SUFFIX alone while TUN is on. The hostname resolves to a synthetic IP first, the rule matches that synthetic IP, and the whitelist never fires. ## 5. Config you can paste Clash/Mihomo syntax. Field names are the same in Clash Verge Rev, FlClash and Hiddify; only the editor location differs. rules: # private ranges must come before any proxy rule and before MATCH - IP-CIDR,192.168.0.0/16,DIRECT,no-resolve - IP-CIDR,10.0.0.0/8,DIRECT,no-resolve - IP-CIDR,172.16.0.0/12,DIRECT,no-resolve - IP-CIDR,127.0.0.0/8,DIRECT,no-resolve - DOMAIN-SUFFIX,lan,DIRECT - DOMAIN-SUFFIX,local,DIRECT - DOMAIN-SUFFIX,home.arpa,DIRECT # add your own corporate suffix, e.g. # - DOMAIN-SUFFIX,corp.example.com,DIRECT dns: enhanced-mode: fake-ip fake-ip-filter: - '*.lan' - '*.local' - '*.home.arpa' - '*.arpa' - '+.corp.example.com' Three notes that save time: - 172.16.0.0/12 covers 172.16 through 172.31. Docker's default networks and several VM subnets live here. Containers that can't reach the host are usually missing this line. - If your rules come from a remote subscription, a local edit gets overwritten on the next update. Put these lines in the client's override/merge section, or in the pre-rules area of the profile. - Reload the profile after editing. Most clients have an explicit reload button; editing the file alone does nothing. ## 6. Verify in order - Gateway reachable: ping succeeds → routing is fine - LAN by IP: open the NAS by IP → rules are working - Resolution: nslookup nas.lan returns 192.168.x.x → fake-ip-filter is working - Hostname: open http://nas.lan → everything is connected Step 2 passes but 3 fails: the problem is DNS. Step 1 already fails: it's TUN routing and rules never ran. Fixing this order into habit beats toggling TUN repeatedly. If LAN works but public access got worse, that's a different problem: see TUN mode connected but no internet for the three-state baseline. To confirm the current exit IP and DNS resolver, use the IP check page. ## FAQ **Q: Disabling TUN restores the LAN but pages get slower. Why?** Without TUN, traffic falls back to the system proxy, so only apps that honour the system proxy are proxied and everything else connects directly. The LAN recovers because the virtual interface is gone. Pages feel slower because apps that never used the system proxy — games, CLI tools — were direct all along, and now nothing looks like it is "using the proxy". **Q: I added IP-CIDR DIRECT rules and it still fails. What now?** Three usual causes: the rules sit below MATCH and never match; the profile was not reloaded; the target is a hostname hijacked by fake-ip, so the rule matched a synthetic address. The four-step check in section 6 identifies which one. **Q: How is Android's "bypass LAN" different from a DIRECT rule?** The system VPN bypass excludes LAN traffic at the OS level, so the client never sees those connections. A DIRECT rule lets the traffic into the client and then releases it. The first is more thorough, the second keeps logs and statistics. They don't conflict. **Q: Our intranet uses a domain that isn't under .lan. Can I use a wildcard?** No. Add the exact suffix twice: DOMAIN-SUFFIX,your.domain,DIRECT in rules, and +.your.domain in fake-ip-filter. Also confirm the client is not forwarding that domain to a public resolver — that returns NXDOMAIN regardless of your rules. **Q: Why can't I reach services running in Docker?** Docker's default bridge networks sit around 172.17.0.0/16, inside 172.16.0.0/12. Without that DIRECT rule, container traffic tries to go out through the proxy and fails. Add the line and reload. ## Related TUN mode connected but no internet — the three-state baseline System proxy on but pages don't load — four checks in order Client download centre — current versions, file sizes and signature verification Error index — match your exact error string to a guide ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-cannot-select-profile-socket-error/ TITLE: cannot select profile: socket error in Clash Verge Rev — HNS check LAST VERIFIED: 2026-11-14 Clash Verge Rev · profile selection # Clash Verge Rev cannot select a profile: socket permission error The subscription is already in the list and the nodes may be fine. The failure happens at selection, inside a Windows service boundary — which is a completely different stage from fetching a subscription, even though both show the word 'profile' on screen. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Read the stage before you act SymptomMore likely stageNext step A profile card will not select, and the log contains An attempt was made to access a socket in a way forbidden by its access permissions verbatimWindows HNS or socket permissionsRecord the error, then check and restart HNS Importing shows client error(Connect)Remote subscription fetchCheck the official import methods, certificates and network Update succeeds but the node list is emptyProfile content, override or parsingCheck the empty-node problem Selection works, but the system proxy leaves you offlineNodes, rules or system proxyMove to the system-proxy page Do not delete profiles, reinstall the client or reset the entire network just because the word "subscription" appears in both the fetch stage and the selection stage. Locate the error to fetch, parse, load, select or traffic capture first. ## Four pieces of evidence before you operate - The full Clash Verge Rev version and the Windows version. - Whether the subscription can still update, whether nodes can be previewed, and whether only selection fails. - The complete error line from the log and its timestamp — with the subscription URL, token, user name and node passwords removed. - The current state of Host Network Service: running or stopped, nothing more. ## When the error matches the official FAQ exactly The Clash Verge Rev Windows FAQ attributes this specific socket-permission error to Windows Host Network Service not running correctly, and gives a restart procedure. - Quit Clash Verge Rev. Do not delete the subscription first. - Open Windows Services and locate Host Network Service. - If it is stopped, start it. If it is running, restart it once. - Reopen Clash Verge Rev and try selecting the existing profile again. - If the error reproduces, keep the error text and the service state and search the official issues with the same version and the same error. The FAQ also provides an administrator command line: net stop hns then net start hns. This site has not reproduced this on every Windows environment and cannot promise that running it restores selection. On a corporate or managed device, contact the administrator instead of working around organisational policy. ## Why the HNS state is worth checking Microsoft documents Host Networking Service as participating in creating network endpoints, addresses, routes and some firewall rules. The official FAQ ties this one socket error to HNS not running, so checking the service state has a direct basis. That is not the same as claiming every socket error comes from HNS — a different error string means a different branch. ## Why redacting the error string matters This branch is defined by an exact sentence. Two errors that both say "socket" but differ in wording can sit in different layers: one inside a Windows service boundary, one inside a subscription HTTP request. Copy the sentence character for character into your notes before you search, and keep the timestamp so you can match it against the log. ## If the error is client error(Connect) That belongs to the remote subscription fetch stage, not selection. The official import documentation describes handling remote subscriptions through a subscription link, a subscription-link profile and URL schemes. For the full HTTP, certificate and content-parsing path, see the subscription update error checklist. The short version: a 401, 403 or timeout means the request never reached the point where a profile can be selected, so HNS is not the first suspect. ## Selected, but the system proxy or TUN still fails - System proxy fails: check the active node, the rules, port conflicts and leftover proxy settings. - TUN fails: check service mode, the virtual adapter, routes and DNS. - Only one website fails: check rule matching, DNS and the site's own status. Continue with system proxy has no internet (Chinese) and no internet after enabling TUN. ## Safe recovery and stopping conditions - Do not run network-repair scripts of unknown origin, delete unknown services or bulk-clean network adapters. - Once the error is reproduced and the evidence is saved, stop mechanical retrying. - If Host Network Service does not exist, will not start, or stops again immediately, hand it to a system administrator or check the Windows event log. - Before asking in public, remove the subscription URL, tokens and private paths. ## Sources and verification date Material checked 2026-10-29. Primary sources: Clash Verge Rev official Windows FAQ — the specific error for a profile that cannot be selected, and the HNS procedure. Clash Verge Rev official subscription import documentation — remote subscription methods and the client error(Connect) boundary. Microsoft Learn, Windows container networking architecture — the role of HNS. This article was assembled with AI assistance and edited by this site against a verifiable boundary; official entries were not expanded into a universal conclusion about all socket errors. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched cannot select profile: socket error. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket error — this pagelocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: Does restarting HNS delete my profiles?** The evidence gathered here is not sufficient to promise that every environment sees no side effects. Back up the configuration you need first, and on a managed device let the administrator handle it. **Q: I cannot find Host Network Service at all. What should I do?** Confirm the service display name and your Windows version first. Do not download a so-called HNS installer from a third party. **Q: HNS is already running, so why can I still not select the profile?** Confirm that your error text matches the official FAQ character for character. A different error should be searched by version and by the stage where it happens. **Q: Can I simply delete the configuration directory?** Not recommended. Deleting it destroys diagnostic evidence and can widen the impact beyond the original problem. **Q: Is this the same as a 401 or 403 on a subscription?** No. 401 and 403 belong to the fetch stage. This specific error occurs in the Windows service or socket permission layer. **Q: What should I include when asking for help publicly?** The client version, Windows version, verbatim error, the stage where it happens and the HNS state — with credentials and private paths removed. ## Continue reading system proxy enabled but no internet (Chinese) · no internet after enabling TUN · the remote profile data is invalid yaml · the full error-string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-remote-profile-invalid-yaml/ TITLE: the remote profile data is invalid yaml — Clash Verge LAST VERIFIED: 2026-11-14 Clash Verge Rev · subscription import # the remote profile data is invalid yaml in Clash Verge Rev This string is often read as "the subscription expired". In the client's own code it marks one specific branch: the HTTP request succeeded, and the response body did not parse as the YAML mapping the client expects. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Which step the error points to In the published parsing code for remote profiles (commit 603e616 at the time of checking), Clash Verge Rev does three things in this order: it checks the HTTP status of the remote request, it decodes the response body as a YAML mapping, and — only after that succeeds — it checks whether the top level contains proxies or proxy-providers.So the remote profile data is invalid yaml means: the body fetched in this attempt did not match what the parser expected. It does not, by itself, tell you that the subscription has expired, that all nodes are dead, or which line is wrong. Code branches move between releases, so this commit describes behaviour at that point, not every historical version. ## Match your evidence to the branch Evidence you haveMost likely branchWhat to do nextfailed to fetch remote profile with status ...The HTTP response failed before parsingWork from the status code and the access conditions; YAML fixes do not applythe remote profile data is invalid yamlBody reached the parser and did not matchCheck the type and structure of what the URL actually returnedError says proxies or proxy-providers is missingYAML parsed, but a required top-level key is absentConfirm you are not importing a different subscription formatImport succeeds but the node list is emptyParsing already passedCheck proxies, proxy-providers, the active profile and proxy-group referencesThese four outcomes come from adjacent branches of the same code path. They are not interchangeable descriptions of one symptom. ## Troubleshoot in order - Keep the working profile. Note the Clash Verge Rev version, the OS, the time of failure and the full error text. Do not delete a profile that still works. Redact the subscription URL, tokens, node addresses and account details before sharing a screenshot. - Confirm the subscription type. In the provider's own dashboard, check that you copied the link intended for Clash/Mihomo. The fact that the same link imports in another client does not prove it is Clash YAML. - Look at what this response actually is. Only on your own trusted machine, check whether the URL returns a configuration, a login page, a verification page, an error page, or nothing at all. Do not paste a token-bearing URL into an online YAML validator, and do not post the full response body to a forum. If it returns HTML or an access restriction, fix the link and authentication in the provider dashboard. If it returns a different subscription format, use the Clash/Mihomo-specific endpoint the provider offers. - If it is YAML, check the structure. After backing up locally, check indentation, colons and quoting, and confirm the top level contains proxies or proxy-providers — the Mihomo configuration reference documents both as configuration keys. A .yaml extension does not guarantee the body is valid YAML, and a file that contains private keys should not be handed to a third-party parsing site. - Change one variable. If the provider offers several formats, switch only to the one labelled Clash/Mihomo and import once. If there is only one link, ask the provider to confirm what format it actually returns. Do not clear the cache, reinstall, change the core and change the subscription in the same attempt — afterwards you will not know which step mattered. - Accept the fix from a new log. Once the error is gone, confirm the active profile switched, that nodes or providers loaded, and that the core starts without a new error. A silent import is not proof that the profile is usable. ## Two easy misreads Base64 that looks like YAML trouble. Issue #2374 records a macOS user whose subscription returned Base64 content; another client accepted it while Clash Verge Rev reported a YAML error. That supports the step "check what format the source returns". It does not prove every occurrence of this error is Base64.Advice from comment threads. Issue #471 records the same error string with a description too thin to yield a general root cause. Do not copy the unreproduced suggestions in that thread — clearing caches, skipping certificate verification, or reinstalling — as if they were a fix.If the message becomes a missing proxies or proxy-providers warning, you have moved from "cannot parse" to "parsed but incomplete". If the update succeeds but the proxy page stays empty, check provider fetching and the use references in proxy groups. Neither of those is still an "invalid yaml" problem. ## Sources and scope Based on the Clash Verge Rev remote-profile parsing code checked on 2026-11-07 (commit 603e616), the Mihomo configuration reference, and public issues #2374 and #471. Public issues are individual reproduction records. This page was not reproduced across every OS, client version and provider, and it cannot see your private subscription response. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched the remote profile data is invalid yaml. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yaml — this pageremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: The same link works in Shadowrocket. Why does Clash Verge Rev fail?** Different clients accept different subscription formats. Confirm that the response is Clash/Mihomo YAML and ask the provider for the matching endpoint. A second client importing it does not prove the response suits Clash Verge Rev. **Q: Does invalid yaml mean the subscription expired?** No. Expiry, a login page, a verification page, an error page and a genuine YAML syntax error all have to be checked separately, both in the returned body and in the provider dashboard. **Q: The file is named .yaml, so why does it still fail?** The extension does not guarantee the body is valid YAML. The actual content may be HTML, Base64, truncated text, or YAML with a syntax error. **Q: Can I paste the subscription into an online validator?** Not advisable. Subscriptions usually embed access credentials. Check on your own device, and send the provider only redacted status, size and format information. **Q: The error is gone after updating, but there are no nodes. What now?** Parsing has passed. Check the active profile, the contents of proxies or proxy-providers, provider fetching, and proxy-group references. That is a different failure stage. **Q: Would disabling certificate verification fix it?** This error gives no evidence that TLS is the problem. Do not turn off certificate checks to work around a format error. Check the request status and the body format first. ## Continue reading set system dns failed: exit code 1 · config.yaml is empty in FlClash · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-service-core-ipc-timeout-retry/ TITLE: core IPC pipe did not become ready within 15s — Clash Verge Rev LAST VERIFIED: 2026-11-14 Clash Verge Rev · core startup # core IPC pipe did not become ready within 15s, then the retry works One red line in the log is not a verdict. The question is whether the core ended up running, and whether traffic actually passes — those are two different checks, and only the second one tells you if the proxy is usable. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## What this page is about On Windows, starting Clash Verge Rev sometimes logs core IPC pipe did not become ready within 15s on the first attempt, while the next attempt succeeds and the proxy works afterwards. The right order is to confirm the final core state first, then deal with the first-attempt warning. The first red line alone does not justify concluding that the service is not installed, that the nodes are broken, or that TUN routing is at fault. This page covers the startup sequence in Windows service mode. If every retry fails, or the service cannot be installed at all, go to service install failed. If the core and the system proxy are fine and only TUN breaks connectivity, go to no internet after enabling TUN. ## Evidence and scope Material checked 2026-11-09 (Beijing time). Official issue #8480 records a Windows 10 x64 case on the 2.5.8 stable client: configuration validation succeeded, the first service core start waited 15 seconds for the IPC pipe and then errored, and the client automatically retried and started the core within about two seconds. The reporter states that the proxy worked after the retry and that no pre-release build was tested. The issue remains a user report: it does not prove that every 2.5.8 installation shares the same root cause, and it is not a confirmed fix announcement. Issue #8474 shows a different state that is easy to confuse with it: the interface says the core has started, but the control API does not respond, the latency test keeps spinning and traffic is not forwarded; restarting the core with the same configuration restores it. That case was Windows 11 in service mode with TUN enabled. "First attempt timed out but the retry succeeded" and "the interface says started but the core is stuck" cannot share one conclusion. This site has not reproduced either case on these system combinations. ## Judge the next step by the result, not the first line What the log and state showWhat it can mean for nowNext step First attempt reports the 15-second IPC timeout, then Service started the core successfully appears and proxy requests workThe first attempt failed but that startup recoveredRecord both attempt times and the version; do not uninstall the service or change nodes The timeout is followed by repeated failures and the core stays downNo usable core baseline yetCheck service state, configuration validation and later errors; move to the service or core-start branch The interface shows the core running, but the latency test spins and /configs or rule-provider requests fail"Process exists" is not the same as a working control API or forwardingSave redacted GUI and core logs; work the stuck-control-API branch Core and system proxy work, only TUN breaks connectivityCore startup is no longer the prime suspectFix nodes and rules, then check the TUN service, interface, routes and DNS ## Check in order — one change at a time 1. Let the automatic retry finish. Look for the next Service started the core successfully after the first error, or for clear repeated failures. Do not keep clicking "restart core" during the retry window; that mixes up the timeline you are trying to read. 2. Confirm a real request. Keep the existing node and rules, disable TUN, use the system proxy to open an ordinary HTTPS page you normally reach, and check whether the client's connection list shows that request. If there is no entry, check whether the system proxy was actually written or whether the application bypasses it. A green interface indicator is not proof that proxying works. 3. Save one timeline. Record the Windows version, the Clash Verge Rev and Mihomo versions, the start mode (Service or Sidecar), how long configuration validation took, the time of the first error, and the time of the later success or failure. Keep only the redacted log lines adjacent to the fault; do not publish the subscription URL, tokens, node passwords or device paths. 4. Separate a cold start from a persistent failure. Quit the client and observe one more startup with the same configuration. If it happened once and the retry recovered, keep the evidence and watch the official issue. If it fails repeatedly, verify that the service runs, that versions match and that the configuration validates — and do not read "Windows cannot find the file" as "the subscription file is gone". 5. Branch only when the symptom matches. If the interface reports success but neither the control API nor traffic works, compare the GUI and core logs against the observation list in issue #8474. If only TUN fails, use the existing TUN decision table. For a proxy that is still working, it is not worth deleting the service directory, clearing system certificates by hand or running repair scripts of unknown origin just to silence one startup warning that already recovered. If you report it upstream, a minimal reproducible sequence and a redacted timeline are far more useful than a screenshot of the dialog. ## What cannot be concluded from here - The 15-second value is the specific log line in that issue, not a fixed threshold for all versions. Wording and retry behaviour can change between builds. - An automatic retry succeeding proves that this particular core start finished. It does not prove that every website, every rule or TUN itself works. - os error 2 appears alongside the IPC pipe not being ready in that log. Those few words alone do not establish that a subscription file, a Geo data file or an antivirus product removed the core. - Issue #8480 does not provide a verified universal fix. If the problem keeps reproducing, keep the original configuration and version and follow the official issue instead of turning one machine's experience into everyone's solution. ## Sources and verification date Material checked 2026-11-09. Primary sources: - Clash Verge Rev official repository, issue #8480 — first service core start waiting 15 seconds for the IPC pipe, then succeeding on retry. - Clash Verge Rev official repository, issue #8474 — interface reports a running core while the control API and forwarding fail. This article was assembled with AI assistance and edited by this site against a verifiable boundary. Neither report was reproduced here, and neither is presented as an official fix. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched core IPC pipe did not become ready within 15s. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15s — this pagefirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: The proxy works after the first error. Do I still need to reinstall the service?** Not yet. Confirm that the following log lines really show the core started, then verify with one real HTTPS request and the client's connection list. If it keeps working, keep the timeline and observe. Only check the service install and version if it fails repeatedly. **Q: The log says Windows cannot find the specified file. Does that mean my profile was deleted?** You cannot infer that. In this case the complete error is that the core IPC pipe did not become ready within the wait window, and the retry then succeeded. Read the full context instead of quoting the second half of a Windows error string. **Q: The interface says the core is running but pages still will not open. Why?** Core process state, control API response, system proxy settings and actual forwarding are separate checks. Confirm the connection list and the system proxy first; if control API requests also fail, work the stuck-core branch rather than switching nodes. **Q: Is this always caused by TUN?** No. The report in issue #8480 happened with both TUN and the system proxy disabled. Only when the core and system proxy are verified working and connectivity fails after TUN is enabled does the TUN branch apply. **Q: Can upgrading to a pre-release build fix it?** There is not enough evidence to recommend that. The original report did not test pre-release builds and the issue has no confirmed universal fix. Before upgrading, read the release notes, back up the configuration and record results under the same conditions. ## Continue reading the service will not install · no internet after enabling TUN · the full error-string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-service-install-failed/ TITLE: Clash Verge Rev service install failed — Is Admin false, program not found LAST VERIFIED: 2026-11-14 Clash Verge Rev · Windows service # Clash Verge Rev keeps asking to install the service and the install fails The client looks identical in all of these cases, but the evidence sits in three different layers: permissions and service files, the virtual adapter, and the proxy path. Mixing them is why people reinstall three times and still see the same prompt. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Decide which of three problems you actually have SymptomLayerRead next The home screen keeps showing Install Service, or the install reports program not foundPermissions, service files, installer flowThis page Running Mode already shows Service, but enabling TUN kills connectivityVirtual adapter, routes, DNSNo internet after enabling TUN Without TUN, the system proxy also cannot reach the internetNodes, rules, system proxySystem proxy has no internet (Chinese) All three use the same window and the same green/grey switch, but they consume different evidence. Reinstalling or resetting the network across all three destroys the one thing you need most: the original error text. ## Save four pieces of evidence before you touch anything - The exact Clash Verge Rev version and the Windows version. - Is Admin and Running Mode as shown in System Info on the home screen. - The complete error text produced by the service install — specifically whether it contains program not found. - The Running Mode, Is Admin value and full version number at the moment of failure. Do not attach a private subscription or node list. The project discussion explicitly recommends judging administrator state from Is Admin: true in System Info, not from remembering that you once right-clicked Run as administrator. That distinction is the whole reason the four items above exist. ## Case 1 — System Info still shows Is Admin: false Official discussion #7475 was filed on Windows 11 with Clash Verge Rev 2.5.1. The reporter stated they had launched the client as administrator, yet the diagnostics still showed Running Mode: Sidecar and Is Admin: false. The accepted answer points out that a genuinely elevated session shows Is Admin: true, and suggests trying a later build of that period and the Service mode. The only conclusion this record supports: do not infer permissions from the act of clicking Run as administrator — read the diagnostics instead. It does not prove that every version needs the same build, and it does not prove that a single UAC confirmation repairs the service. ## Case 2 — the error text contains program not found Official issue #8138 documents a Windows 11, v2.5.7+autobuild case: installing the service in normal user mode produced program not found. The reporter then fully exited the client, started it as administrator, saw administrator mode confirmed in System Info, and reports that the service installed afterwards. - Exit the client completely — not just close the window — then start it elevated and confirm the UAC prompt. - Return to System Info and check that Is Admin: true is actually displayed. - Attempt the service install once more from that state. - If it still fails, keep the version, Windows build, Running Mode and verbatim error. Do not try to guess which file is "missing" from the words program not found. - Take installers only from the official Releases page. The reporter in #7475 later mentioned an old-version conflict remaining after an update; that is a single record, which is why this page does not present "reinstall" as a standard fix. ## Case 3 — installation finished without satisfying the requested service and cores Official issue #8023 was filed on Windows 10 with v2.5.4 and v2.5.5. The error was installation finished without satisfying the requested service and cores, and the log also contained a core path security check. The report establishes only that this error appeared in those versions and that environment. - First confirm that your error text and version match the issue. A different error must not inherit the same treatment. - Do not copy the ACL modifications, service deletions or system-directory cleanup described in the thread. Path permissions are a security boundary; changing them incorrectly widens the exposure. - Check the official Releases page and same-version issues, verify with one later stable build, and test one version at a time while recording Running Mode. - If service mode is now healthy and only TUN traffic fails, stop reinstalling and move to the route and DNS checks on the TUN page. ## Case 4 — service or core errors appear only after an upgrade Service binaries and core files change between releases. Check the official Releases page for the current stable version and known issues before deciding to upgrade or roll back, and do not overwrite-install across several versions in a row. - Export the configuration you need beforehand, and keep the subscription URL in a separate, private place rather than in a public log. - Test one version at a time and record the install result plus Running Mode after each attempt. - An issue report proves that a specific environment produced a specific error; it does not mean every user should pin the same old build. When the Releases page was checked on 2026-10-28 it showed v2.5.8 with a date of 2026-10-09. Versions keep moving, so re-check the official page before acting on this paragraph. ## When to stop retrying If you have saved the full version, Running Mode, Is Admin value and verbatim error, and the same operation still reproduces the failure, stop. Do not keep editing ACLs, deleting system services or running scripts from strangers. Search the official issues with the identical version and identical error string; on a managed device, contact the administrator. This stopping rule is risk control, not a repair effect verified in any issue. ## What this page cannot promise Official issues and discussions are evidence of a problem, not a blanket repair commitment across Windows versions. This page has not been tested on every antivirus product, managed device or legacy Windows build, and it will not recommend modifying unknown ACLs, deleting system services or running unfamiliar scripts. If an error names a specific path, redact the user name, device name and subscription details before sharing it. ## Sources and verification date Material checked 2026-10-28. Primary sources: Clash Verge Rev official Releases — current release and version notes. Official discussion #7475 — using System Info to read Is Admin and Service mode. Official issue #8138 — reproduction record for program not found during service install. Official issue #8023 — individual case of installer and service/core state not matching. This article was assembled with AI assistance and edited against a verifiable-evidence boundary; issue comments were not rewritten into an official universal fix. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched service install failed / Is Admin false / program not found. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not found — this pageservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: Why does enabling TUN keep prompting me to install the service?** Look at Running Mode in System Info first. If it still says User, the client is not running in service mode; then check Is Admin, the full version number and the verbatim install error. TUN needs the service, so the prompt is the symptom, not the cause. **Q: If Is Admin is true, will the service install succeed?** Not necessarily. Is Admin: true only confirms that the diagnostics report an elevated session. It does not guarantee the install succeeds and does not rule out every permission problem; search the issues with the identical error and version. **Q: Should I download a separate service file when I see program not found?** Not from an unknown source. Use the official complete installer only. The error string does not by itself prove that a particular file is missing, so check Is Admin, Running Mode, version and the full error text before assuming anything. **Q: Does issue #8023 apply to every version?** No. That issue records Windows 10 with v2.5.4 and v2.5.5 and one specific error string. A different version, error or log means the cause described there cannot simply be transferred. **Q: The service installed but TUN still has no internet. What now?** Confirm in System Info that Running Mode is already Service. Only when service mode is healthy and connectivity still fails with TUN enabled should you move to the TUN page and check the virtual adapter, routes and DNS. Those are investigation directions, not causes you can confirm from the symptom alone. **Q: Can I paste the complete log into a public issue?** Redact it first. Remove the subscription URL, tokens, node passwords, user name, device name, internal addresses and identifiable paths, keeping only the version, timestamps and the necessary error lines. ## Continue reading no internet after enabling TUN · core IPC pipe did not become ready within 15s · cannot select profile: socket error · the full error-string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-set-system-dns-failed-exit-code-1/ TITLE: set system dns failed: exit code 1 — Clash Verge Rev LAST VERIFIED: 2026-11-14 Clash Verge Rev · macOS / Windows / Linux # set system dns failed: exit code 1 in Clash Verge Rev Configuration validates, the core starts, and then the client fails on the step that writes the system DNS. Here is how to find out which stage broke, and what the public evidence does and does not support. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## What the error actually tells you set system dns failed: exit code 1 means Clash Verge Rev did not succeed at the step where it modifies the system DNS configuration. It does not prove that the subscription, the nodes or the Mihomo configuration are broken, and it does not follow that swapping in 1.1.1.1 will fix it.Two things have to be established before you change any setting: whether the failure sits in the system proxy path, in the TUN path, or in basic networking underneath both; and whether another tool that also takes over networking is running at the same time.The one public case that names this error is Clash Verge Rev issue #8328: macOS 27.0.1, Clash Verge Rev 2.5.8, configuration validation succeeded, and the error appeared when enabling the system proxy. The reporter then pointed at a Tailscale/WireGuard type conflict, and a project collaborator noted that Mihomo supports WireGuard and Tailscale outbounds directly. The issue was closed on 2026-10-17. It is one case on one platform — it does not establish that every error with this string comes from Tailscale. ## Route by symptom before anything else What you seeMore likely stageDo this firstDo not do thisValidation succeeds, then set system dns failed: exit code 1System DNS write, or a network-takeover conflictRecord OS, client version, and every VPN or virtual adapter runningDo not delete the subscriptionSystem proxy fails but TUN still worksSystem proxy and DNS write pathTurn TUN off, build a direct baseline, reproduce system proxy aloneDo not treat working TUN as proof that DNS recoveredNo connectivity only after enabling TUNTUN service, routing, or DNS takeoverSwitch to a TUN-specific checklistDo not apply the system-proxy conclusions hereName resolution fails even with direct connectionOS, router, ISP or managed-network DNSRestore ordinary networking firstDo not stack more client settings on topOnly one site will not openSite-specific DNS, certificate, region or account policyCompare against another domain and a resolver queryDo not reset the whole device network ## Step 1 — Capture the first error and the reproduction conditions Record these before you touch anything:- Full Clash Verge Rev version, OS name and version. - Whether the error appears when enabling the system proxy, enabling TUN, switching profiles, or starting the app. - Whether configuration validation succeeded, and the first ERROR line before the failure. - Whether Tailscale, WireGuard, another VPN, a network filter or corporate security software is running. - Whether ordinary networking and name resolution come back after you close Clash Verge Rev. Redact subscription URLs, tokens, node addresses, UUIDs, usernames, device names and your real public IP before posting a log or a screenshot. Do not upload a full configuration file just to report an error. ## Step 2 — Build a baseline that does not pass through the proxy - Turn off both TUN and the system proxy in Clash Verge Rev. - Quit the client normally. Do not leave a second proxy client running. - Before pausing another VPN or virtual network tool, confirm what it is for and how to bring it back. On a company-managed device, do not remove it yourself. - Confirm that ordinary HTTPS sites and name resolution work on a direct connection. - Reopen Clash Verge Rev and enable only the system proxy. Watch whether the same error returns. If name resolution is broken even on a direct connection, fix basic networking first; this page does not apply yet. Avoid one-click network-reset scripts of unknown origin, and do not delete system network services blindly. ## Step 3 — Look for duplicate network takeover Issue #8328 points at Tailscale/WireGuard type tools, but that is single-case evidence. Check each of these:- Is another VPN connection still listed in the menu bar or system settings? - Are Tailscale, WireGuard, ZeroTier, a corporate VPN, a packet filter or security software installed? - Is a background service or network extension from an older proxy client still present? - Are Wi-Fi, Ethernet and virtual interfaces being given DNS settings by different programs at the same time? Pause one known tool at a time, then repeat the same action. Once a conflict is confirmed, decide which tool you actually need — do not uninstall a corporate VPN that your work depends on just to make Clash Verge Rev start. WireGuard and Tailscale outbounds inside Mihomo are an advanced setup; never paste a configuration you found online into a private subscription. ## Step 4 — Test system proxy, TUN and DNS separately The system proxy and TUN are different takeover paths, so they have to be tested in isolation:- Enable only the system proxy. Check that the core is running and that the browser honours the system proxy. - Then test TUN on its own. Do not change the subscription, the rules and the DNS settings in the same attempt. If the system proxy throws set system dns failed while TUN works, still record that the system DNS write failed. Do not write it up as "resolved".If the system proxy enables cleanly but names still fail, compare the OS resolver output with the client log, then move to a general DNS checklist.If only TUN fails, use a TUN-specific checklist instead of this page.The Mihomo DNS documentation lists separate fields for default-nameserver, nameserver, proxy-server-nameserver, respect-rules and Fake-IP. Their existence shows that the DNS chain has several layers; it does not prove that any one field is the cause of your exit code 1. If you do not know the override order of your subscription, do not replace an entire DNS block in one edit. ## Step 5 — Safe recovery and stop conditions Change one variable per attempt, and keep the error text from before and after.To exit: disable TUN first, then the system proxy, then quit the client normally.Restore the system's original automatic DNS or the administrator-assigned settings, then verify a direct connection.Install the client only from the official Clash Verge Rev project and its Releases page.If the error reproduces reliably, report OS version, client version, the action that triggers it, the first error line, and a redacted log.Stop and escalate if the device is managed, if an unknown network extension is involved, or if administrator rights are needed. ## Sources and evidence boundary Clash Verge Rev issue #8328 — one macOS 27.0.1 / 2.5.8 case; validation succeeded, then the system DNS write failed, and the discussion points at Tailscale/WireGuard type conflicts. Closed 2026-10-17.Mihomo DNS configuration reference — used only to show that the DNS configuration has multiple layers, not to claim that one field fixes every exit code 1.Clash Verge Rev official project — used to check the project, issue and release entry points.This page assembles a diagnostic order from public material. It was not reproduced across every macOS, Windows, Linux, VPN and network-extension combination, and it does not promote a single closed issue into a universal cause. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched set system dns failed: exit code 1. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1 — this pageDNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: Does this error mean my nodes are broken?** No. In issue #8328 the system DNS error appears only after configuration validation succeeds, so the failure happens earlier than any judgement about node health. Test the nodes separately once connectivity is restored. **Q: Can I just set the DNS to 1.1.1.1 or 8.8.8.8?** Not as a general answer. The conflict may sit in a system network extension, another VPN, the TUN path, or the client's override layer. Swapping resolvers can mask the problem and may also break a corporate network requirement. **Q: TUN works, so do I still need to fix the system proxy error?** Yes — record it and treat it separately. Working TUN only shows that a different takeover path happens to function; it does not mean the system DNS write succeeded. **Q: Must I uninstall Tailscale or WireGuard?** Not necessarily. First confirm that the conflict reproduces and decide which tool you actually need. On a managed device or a work VPN, hand this to your administrator instead of uninstalling based on one public case. **Q: Will reinstalling Clash Verge Rev fix it?** There is no evidence that guarantees it. If the conflict comes from another network extension or from system policy, reinstalling the client does not remove the conflict. Keep the log and finish a single-variable comparison first. **Q: What should I include when I open an issue?** OS version, Clash Verge Rev version, the exact action that triggers it, the configuration validation result, the first error line, whether another VPN was running, and the minimum redacted log. Never include subscription URLs, node credentials or your real IP. ## Continue reading Invalid YAML when importing a remote profile · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-system-proxy-no-internet/ TITLE: Clash Verge Rev system proxy on but no internet — the four parts to check in order LAST VERIFIED: 2026-11-17 Clash Verge Rev · System proxy # System proxy is on but there is no internet in Clash Verge Rev When the system proxy is enabled and pages still fail, the order matters: get plain networking back first, then check the core, the local port the system proxy points at, the node, and only then rules and DNS. A green “configuration validated” message does not mean the proxy is actually carrying traffic. Node Handbook editorial team · verified 2026-11-17Material checked on 2026-11-16. This page covers system-proxy mode only. If the failure starts when you enable TUN, use the TUN page instead. If you are not sure which takeover method you are running, read the error index first and come back. ## Short answer Disable system proxy and TUN, quit the client completely, and confirm ordinary browsing returns. Then re-enable only the system proxy and check four things as separate items — core running, local port matches the current config, node has current connectivity evidence, rules and DNS match the failing request. Change one item per round and keep the result. Most “it broke after the update” cases turn out to be a port or takeover-mode mismatch, not a dead node. ## Separate the proxy mode from the system proxy switch Selecting a mode inside the client does not mean the operating system has handed web requests to Clash Verge Rev. Keep TUN off and look at three places, each of which answers a different question: the client's system-proxy switch, the browser's own proxy setting, and the client's connection log. What you seeWhat it provesNext step A browser with the local port typed in manually can load pages; other apps cannotThe local port and that browser's path may work; it does not prove the system proxy is enabledCompare the client switch with the OS proxy settings, then retest the same URL The system proxy switch is on, but the client shows no matching connectionThe request may never have reached the current coreCheck the address and port the system proxy points at, and whether the browser overrides it The client logs a failed connectionThe request did reach the core; the fault is in the node, rules or DNSKeep that record and continue with step 4 This is a diagnostic order, not a verdict. Some applications ignore the OS proxy setting entirely, and browser extensions can override it. Only ever compare the same URL, the same node and the same takeover method. ## Step 1 — turn takeover off and confirm plain networking returns Switch off both the system proxy and TUN, quit the client, then test ordinary browsing. If networking is still broken after quitting, open the OS network proxy settings and check whether a local proxy address is still configured there; if the switch is still on, turn it off and retest. ResultWhat it meansNext Plain networking returns after disabling takeover and quittingThe core path is the suspect, not the OSContinue to step 2, enabling only the system proxy The OS still holds a local proxy addressThe switch state and the written setting disagreeTurn the remaining switch off and retest before anything else Networking is still broken with takeover fully offThis is no longer a proxy-client questionStop toggling the proxy and work through a general connectivity checklist first Do not install unknown certificates or disable TLS verification in order to get back online. ## Step 2 — enable only the system proxy and verify four parts separately Reopen the client, enable the system proxy, and keep TUN off so that two takeover methods are never in play at once. The mihomo documentation treats proxy ports, DNS and rules as separate concerns, and the diagnosis has to do the same: a running core does not mean the system proxy points at it, a correct address does not mean the node is reachable, and neither of those says anything about rules or DNS. PartWhat to checkWhat it does not prove Core stateIs the current core running, and is there a log line or connection record for this requestThat configuration validation equals a working path System proxy and local portThe local address and port written into the OS settings, against the proxy port in the current configThat the switch being on means it points at the right place Node connectivityWhether the selected node has current evidence tied to this nodeThat an imported subscription means a usable node Rules and DNSWhether the failing request matched the rule you expected, and whether DNS errors appear with itThat one failed request condemns every node or the whole config Change one item at a time and write down the before and after. If you swap config, node and takeover method in the same round, a working result tells you nothing about which change fixed it. ## Step 3 — check the port the system proxy actually points at Read the local address and port recorded in the OS proxy settings, then compare them with the config currently in use. This page deliberately does not hand out a fixed port to copy: the correct value is whatever the current config exposes. Record four things — the address and port in the OS settings, the proxy port in the config, whether they match, and whether the core is running. If they disagree, fix that first and repeat the same request. If they agree and it still fails, move on to the node, rules and DNS; “the ports match” is not a clean bill of health. ## Step 4 — node evidence first, then rules and DNS Node source, import result and connectivity result are three different records. For any failing request, capture: the target URL and the exact symptom, the node in use, the rule that matched, whether a DNS error appeared alongside it, and whether the result changed after a single adjustment. Without those five items the honest conclusion is “not yet localised”, not “it's DNS” and not “the node is dead”. ## If the same proxy works in another client, is Clash Verge Rev broken? It is useful comparison data, and it is not a diagnosis. Official issue #668 records a case where networking failed with the system proxy enabled while the same proxy worked in another client; the reply notes that a Clash for Windows configuration is not necessarily compatible with the mihomo core. That makes config compatibility a legitimate item to check — it does not make it the common cause of every similar report. ## Why can networking stay broken after the proxy is switched off? Configuration validation and proxy restoration are separate states. Issue #4527 records a Windows 11 user who still had no connectivity after turning the system proxy and TUN off, with domain resolution still returning Fake-IP addresses; the original post shows an empty log pane, while a comment supplies a log with successful validation, and the reporter says upgrading to 2.4.1 resolved it. Discussion #5822 contains a validated config plus a note that a browser pointed manually at the local port behaved differently from other apps. Both cases point at the same discipline: treat validation, the client switch, the OS proxy state and the actual request result as four separate facts. These are individual reports from specific versions and devices. They are not a statement about all versions or all users. ## One URL, one explicit proxy request When the system proxy is on but the client shows no matching connection, take the port from your current config and make one explicit request on the same device and the same URL: # HTTP or mixed port curl --proxy http://127.0.0.1:YOUR_PORT -I https://example.com # SOCKS port only curl --socks5-hostname 127.0.0.1:YOUR_PORT -I https://example.com Replace YOUR_PORT with the value from your own config; do not paste a number from someone else's guide. Read the result as a fork: system proxy fails but the explicit request succeeds → check what the OS actually wrote (address, port, active network service). Both fail → go back to core logs, node and rules. Both succeed but one application fails → that application does not follow the system proxy. Record the test time, OS version, client version, port type, response code and the matching connection entry. Do not post subscription URLs or tokens in screenshots. ## Record sheet ItemThis round Plain networking restored after disabling takeover and quittingto fill OS still holds a local proxy addressto fill Only the system proxy enabled during this roundto fill Core running, matching record presentto fill System proxy address/port equals the current configto fill Selected node has current connectivity evidenceto fill Rule and DNS record for the failing requestto fill Browser has its own manual proxyto fill Single change made and its resultto fill ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched system proxy is on but no page loads. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loads — this pagesystem proxy written, traffic not routed TUN mode is on but apps still bypass itvirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and the network first, then come back with the exact log line. Changing several switches at once destroys the evidence you need to tell them apart. ## Frequently asked questions **Q: Does turning on the system proxy mean all apps are proxied?** No. The system proxy only writes an address and port into the OS settings, and applications that ignore those settings will still connect directly. TUN captures traffic at the IP layer instead, which is a different mechanism with different failure points. **Q: Configuration validation passes, but nothing loads. What does that mean?** It means the config file parsed and the core accepted it. It says nothing about the port the OS is pointed at, whether the selected node is reachable, or whether the rules match your request. Check those three separately. **Q: The ports match and it still does not work. What next?** Move to node evidence: does the client log a connection attempt for this request at all? If it does and the connection fails, the fault is at the node, the rules or DNS. If it does not, the request never reached the core, so the problem is still in the takeover path. **Q: Should I just reinstall the client?** Not first. Reinstalling discards the evidence that would tell you which layer failed. Restore plain networking, capture one failing request with its log line, and only then consider a clean reinstall as a second experiment. **Q: Does this apply to Clash Verge Rev 2.x only?** The layered order applies to any mihomo-based client. The specific menu names and the issue numbers quoted here are tied to the versions in those reports, so cite the version when you reuse a conclusion. ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/clash-verge-rev-tun-not-working/ TITLE: Clash Verge Rev TUN mode not working: connected, no internet LAST VERIFIED: 2026-11-14 Clash Verge Rev · TUN # TUN mode connects but there is no internet in Clash Verge Rev If the system proxy also fails, TUN is not your first problem. This page only applies once the proxy path is known to work and enabling TUN alone breaks connectivity — and it is built around one-node, one-rule comparisons. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## First decide whether this is really a TUN problem - Disable TUN, keeping the current subscription, mode and node. - Enable the system proxy and open an ordinary HTTPS site. - Confirm a connection appears in the client's connection list, then compare the exit IP with an IP check page. If the system proxy also fails, work on the subscription, nodes, core or system proxy first. Only when the system proxy is fine and enabling TUN alone kills connectivity does the rest of this page apply. The official quick start also treats the system proxy and TUN as two different capture methods: the system proxy mainly covers programs that honour system settings, while TUN captures games, command-line tools and other software that never reads the system proxy. Scope: Clash Verge Rev desktop on Windows, macOS and Linux. Interface labels can change between versions. This page is a reproducible diagnostic order, not a claim to cover every system, antivirus product or corporate network policy. ## Build a three-state baseline on one node and one rule set To avoid changing node, mode and the TUN switch at the same time, first record the OS version plus the Clash Verge Rev and Mihomo versions, then pin one node and one rule mode. Use a plain website that your network normally reaches as the direct baseline, and pick a separate HTTPS target you expect to reach through the proxy. Change only the capture method at each step. StateWhat to recordHow to read the result System proxy off, TUN offA site you know is reachable directly, plus system network and DNS stateIf a single site fails, try another known-reachable target before declaring a base-network fault Same node, system proxy on, TUN offProxy test page, client connection list, exit IPA failure here points at the subscription, node, core or system proxy System proxy off, TUN onlyKeep node, rules and target unchanged; check that the test application has no separate proxy setting, then check the virtual interface, connection list, page load and DNS resultOnly a matching connection entry after the system proxy is off justifies looking further into TUN capture; if only this step fails, check the service, routes and DNS capture Keep the timestamp of each step together with the redacted log lines. If the service install already reported an error, move to service install failed; if the system proxy itself is broken, move to system proxy has no internet (Chinese). Neither situation should be treated with TUN settings first. ## Step 1 — service mode and permissions TUN has to create a virtual adapter and adjust system routes, which needs more privilege than an ordinary system proxy. Confirm in the settings that service mode is available before enabling TUN. After installing or updating the service, restart the client, and do not run an old Clash build, another VPN or a second Mihomo frontend at the same time. - Windows: confirm the service install reported no error and that Windows Security or a third-party firewall is not blocking the Mihomo core. - macOS: check the system network and firewall settings; if the log clearly reports a permission denial, follow the operating system's own prompts and record the result before and after any change. - Linux: check whether the current service log reports a TUN device or route write failure; permission configuration differs by distribution and should follow that system's documentation. Do not rely permanently on "start as administrator every time" to paper over a failed service install. Permission only allows TUN to be established; it does not make nodes, DNS and routes correct. ## Step 2 — did the virtual adapter actually appear? After enabling TUN, look at the system adapter list: an interface created by Mihomo or Clash should normally be present. If the switch flips straight back, the log reports a creation failure, or no new interface appears at all, keep checking the service, permissions and driver. The interface list alone is not enough to identify a single cause. If the interface exists but all traffic drops immediately, continue with routes and DNS. Do not hand-delete physical adapters you do not recognise; when cleaning up, only target virtual interfaces confirmed to belong to this client, and quit other proxy software first. ## Step 3 — default route and multi-adapter conflicts The Mihomo documentation states that auto-route routes system traffic into the TUN device automatically, and auto-detect-interface identifies the actual outbound interface. A laptop connected to Wi-Fi, wired Ethernet, a phone hotspot, a virtual machine, Docker and a work VPN at the same time can hit conflicts in that automatic choice. - Temporarily disconnect unused VPNs, VM adapters and hotspots, keeping one real uplink. - Disable TUN, confirm direct connectivity returns, then enable it again. - If only the corporate intranet, a NAS or LAN devices become unreachable, check whether those private ranges need to be excluded — do not switch all traffic to direct. - If the fault appears only after moving from Wi-Fi to wired, fully disable and re-enable TUN so routes are rebuilt. Do not copy a stranger's entire TUN configuration block. Route exclusions depend on your own subnets; copying them can leak intranet traffic or permanently bypass the proxy for a target address. ## Step 4 — separate "no network" from "DNS failure" If IP addresses connect but domain names do not, or the log shows resolution timeouts, NXDOMAIN or SERVFAIL, move to DNS. Mihomo's dns-hijack sends matching DNS connections into its internal DNS module, and Windows and macOS place limits on automatically hijacking DNS requests aimed at the local network. - Turn off the browser's secure DNS or another standalone DoH resolver once as a comparison — not as a permanent downgrade of security settings. - Only on devices you administer, temporarily compare a personal DNS filter such as AdGuard, then restore it after the test. For organisation-managed security clients or antivirus network filtering, contact the administrator instead of disabling them. - Confirm the DNS module is enabled in the configuration and that the upstream resolvers are reachable from the current network. - Adjust Fake-IP, IPv6 or strict-route only when you know what they do, one item at a time, keeping the original configuration. ## Step 5 — use the log to assign the fault to a layer SymptomCheck firstNext step The TUN switch flips backService, permissions, core startSave the verbatim error, check the official version and service state Interface exists, everything is offlineDefault route, other VPNs, multiple adaptersKeep one uplink, rebuild TUN IP works, domain failsDNS, DoH, Fake-IPCompare system DNS with client DNS Browser fine, game failsUDP, rules, node capabilityCheck whether that application's connections reach the client Only intranet or NAS unreachablePrivate routes and exclusionsConfigure against your real subnet, never a copied template When asking for help, provide the OS version, Clash Verge Rev version, Mihomo core version, time of the fault and redacted logs. Do not publish the subscription URL, tokens, node passwords, device names or internal addresses. ## How to restore networking safely - Disable TUN inside the client. - Disable the system proxy and service mode, then quit the client normally. Confirm no system proxy remains, and reconnect Wi-Fi or Ethernet. If connectivity still fails, record the current adapter and route state and reboot as a comparison. A reboot does not guarantee removal of the root cause, and a full network reset should not be the first move. After recovery, rebuild the baseline from "system proxy works" and then enable TUN again — that tells you which step reintroduced the fault. ## When to stop changing configuration If the subscription is expired or the provider is doing maintenance, fix that first and retest TUN afterwards. A failing system proxy only means a usable baseline has not been established: follow the recovery steps, confirm TUN is off and any leftover routes or DNS effects are gone, then check the base network, subscription and system proxy. If the problem happens only on a corporate or campus network, network policy may be involved; do not work around organisational restrictions without authorisation. For core crashes or repeatable service install errors, search the official GitHub issues with the same version after redacting your logs. ## Keep a before/after route comparison on the same device When the system proxy mode works and TUN mode breaks connectivity immediately, the most useful artifact is the route difference on the same device, same network and same node. On Windows, route print shows the interface list plus the IPv4 and IPv6 tables; macOS and Linux have their own built-in route commands. Do not copy another device's default gateway or subnet configuration. With TUN off, record the time, the active adapter, the default route and the result of one ordinary HTTPS request. With TUN on, first confirm the core and virtual adapter actually started, then record the same items. If TUN was never created, a route comparison cannot prove a route conflict. Compare the default route, the target intranet subnet and interface priority. Only when the route clearly changes and the failure appears at the same moment should you examine exclusion rules further; if the route is unchanged and only domains fail while IP requests succeed, go to DNS first. Restore the original settings and retest the same site to confirm normal networking returned. Mask internal addresses, device names and subscription details when publishing logs. Field meanings for the Windows command are documented by Microsoft's route reference. This comparison locates the failing layer only; it cannot by itself prove that a node stays stable over time. ## Sources and update record Material checked 2026-11-05. Primary sources: the Clash Verge Rev official quick start, the official project repository, and the Mihomo documentation for TUN configuration. All were reachable at the time of checking. The three-state table is this site's diagnostic method, not a measured benchmark. This article was assembled with AI assistance and edited by this site against a verifiable boundary: forum cases were not rewritten into universal conclusions, and no device testing was invented. ## Which of the seven Clash Verge Rev failures you actually have You are on this page because the symptom matched TUN mode is on but apps still bypass it. Clash Verge Rev reports six other failures that beginners routinely mix up with it. Match the exact string in your log or UI against the table before you change any setting — each row leads to a different fix, and two of them are not configuration problems at all. What the log or UI saysWhat it actually points to system proxy is on but no page loadssystem proxy written, traffic not routed TUN mode is on but apps still bypass it — this pagevirtual adapter / route layer service install failed / Is Admin false / program not foundservice never installed core IPC pipe did not become ready within 15sfirst service-core start times out, retry works cannot select profile: socket errorlocal socket permission / HNS state the remote profile data is invalid yamlremote fetch returned something the parser rejects set system dns failed: exit code 1DNS write stage failed after the core started If none of the seven matches, the failure is below the client: check the node and th ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/flclash-config-yaml-empty-profile-switch/ TITLE: config.yaml is empty — FlClash profile switch fails LAST VERIFIED: 2026-11-14 FlClash · configuration profile # config.yaml is empty when switching profiles in FlClash The message appears the moment you switch profiles, and the node list may look perfectly healthy. That combination matters: it puts the failure at configuration generation and loading, not at the subscription. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Judge by the stage first SymptomMore likely stageDo firstDo not do firstconfig.yaml is empty immediately on switching — consistent with issue #1784, but still check your version and current profileConfiguration switchRecord version, profile name and exact error textDo not wipe app data straight awayNode list empty after a subscription updateFetch, response or parsingMove to a subscription update checklistDo not just keep re-testing latencyNode list fine, error only when starting the coreLoading or runtimeSave the core log and work by stageDo not apply this page's conclusionAppears on an old build, gone after updatingPossibly an already-fixed pathKeep a before/after comparisonDo not generalise to all empty-config problems ## Step 1 — Keep a restore point and get ordinary networking back Record the FlClash version, the OS version, the time of the error, the action that triggered it, and whether the profile is still listed. A full subscription URL usually carries a token — keep it out of public screenshots, issues and unredacted logs.If ordinary networking is also affected, disable the system proxy or TUN and confirm whether plain HTTPS sites load without the proxy. This reduces the number of variables; the available evidence does not show that it repairs an empty configuration error. ## Step 2 — Check the version and the scope of the official fix - Below v0.8.97: back up the configuration and any override copies first, then check the official Releases for an upgrade. - v0.8.97 or newer: do not assume it is still the same defect. Continue with the current profile and the subscription response. - Crossing v0.9.0: the official release notes state a breaking change — custom overrides were removed and must be rebuilt as a custom profile. - After updating, test the base configuration before restoring TUN, WebDAV and custom settings one by one. A version number narrows the field. It does not replace the log and the current profile state. ## Step 3 — Confirm the current profile still exists - Go back to the profiles screen and check whether the selected profile is still listed. - Keep the original profile for now; do not delete, reinstall and clear data in the same attempt. - If you need to re-add it, create an independent test profile first. - Record whether the test profile can be added, and whether switching to it produces the same error. - Only remove the old profile once the comparison is stable. A new profile switching successfully proves that the two profiles behave differently. On its own it does not establish why the old one fails. ## Step 4 — Separate a successful fetch from a generated configuration If the problem occurs while fetching the subscription rather than while switching, use a subscription update checklist instead. A browser receiving HTTP 200 does not mean the client has parsed the payload, written it to disk and loaded it.Record the HTTP status, the response type and whether the body is empty, without publishing the body. Connection reset by peer is a separate case at the fetch stage and cannot be explained by an empty-configuration fix. If the error appears only after content is retrieved, keep checking the format, the current profile, and the configuration generation stage. ## Step 5 — Restore advanced settings one at a time Once the base configuration switches correctly, restore one variable per attempt: custom profile, system proxy, TUN, WebDAV. Re-check the current profile, proxy groups and ordinary networking after each one. If the error returns after restoring a particular item, stop stacking changes and keep the before/after difference.If names fail to resolve while raw IPs still work, move to DNS troubleshooting — that link does not imply DNS is the cause of an empty config error. ## Sources and scope Material checked 2026-10-31 against FlClash issue #1784, the official FlClash Releases page, and issue #1284 as a separate fetch-stage case. This page was not reproduced on every platform and version. Official issues describe individual environments, and they do not guarantee that an upgrade resolves every empty-configuration problem. ## Frequently asked questions **Q: Does 'config.yaml is empty' mean the nodes are dead?** No. The available evidence concerns profile switching and an empty configuration being written. It says nothing about node connectivity. **Q: Will upgrading to v0.8.97 fix it?** Not guaranteed. That version fixed one confirmed path; other causes still have to be checked separately. **Q: Can I delete the FlClash data directory?** Not as a first step. It removes the configuration, the overrides and the diagnostic trail. **Q: The browser opens the subscription, so why does FlClash fail?** The browser only proves a response was returned. The client still has to parse it, generate a configuration and load it. **Q: Is 'Connection reset by peer' the same problem?** No. It is a distinct error recorded at the fetch stage and should not be merged with this one. **Q: What should I include in an issue?** Version, OS, the stage that triggers it, the shortest reproduction steps, and the redacted error text. Do not attach a full subscription or node details. ## Continue reading the remote profile data is invalid yaml · ArgumentOutOfRange_Count when updating subscriptions · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/hiddify-warp-parseprefix-empty-ip/ TITLE: Hiddify WARP: netip.ParsePrefix error, empty IP LAST VERIFIED: 2026-11-14 Hiddify · WireGuard / WARP outbound # netip.ParsePrefix("/24"): unable to parse IP in Hiddify WARP The address prefix has nothing in front of the slash. That is a syntax-level fact about the generated data — not a verdict on your subscription, your carrier, your WARP account, or the client as a whole. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## What the string means The full error usually reads error patching outbound ... netip.ParsePrefix("/24"): ParseAddr(""): unable to parse IP. It is raised while the WARP outbound configuration is being generated or parsed. The /24 with no address before it means the data being handed to the parser is not a complete address prefix.What it cannot tell you on its own is which party produced the empty value: the subscription, the carrier, the WARP account, or the client. Establish first whether WARP is the only broken path, or whether the rest of the configuration also fails.Three public reports name this family of error. Issue #930 is Ubuntu 22.04, at the point of adding WARP. Issue #1389 is Windows with Hiddify v2.0.5, after the WARP configuration was added and a connection was attempted. Issue #1687 is Linux on Hiddify 2.5.7, during outbound patching. Together they show the fault is not confined to one platform; none of them establishes a fix that applies to every version. ## Judge by the stage where it appears What you seeStage to judgeNext actionParsePrefix("/24") appears the moment you add or generate WARPThe generated WARP address field, or local configuration stateKeep a redacted log; check the Hiddify version and the action that generated itWARP entry appears, error only on connectOutbound configuration loadingDisable WARP and test the original subscription aloneNo WARP involved, node list empty after a subscription updateSubscription fetch, parse or selectionUse a "no nodes after update" checklist insteadOriginal nodes work once WARP is offWARP path differs from the ordinary node pathKeep the working configuration; do not substitute reinstalling for diagnosisDirect connection also fails with WARP offBase networking, or system proxy/VPN stateRestore ordinary networking before testing the client again ## Step 1 — Keep a recoverable baseline - Record the OS, the full Hiddify version, where the configuration came from, and whether the error appeared at "generate WARP" or at "connect". - Record the first relevant log line before the error, with the subscription URL, WARP key, node addresses, username and public IP redacted. - Disable WARP, select a profile you know worked, and verify that ordinary HTTPS pages and your egress IP come back. - If ordinary networking still fails, quit Hiddify and check the system VPN or proxy state before generating another WARP configuration. This step builds a comparison point. It does not prove a single root cause. Do not delete the whole configuration or upload an unredacted log. ## Step 2 — Check the WARP configuration through the official flow The Hiddify WARP guide separates the built-in feature from a custom warp:// configuration. First confirm which of the two you are using, then check whether the custom configuration contains an empty address or an incomplete prefix. Button positions vary between versions, so follow the interface of the build you actually run.The sing-box documentation for the WireGuard address field requires a list of IP addresses with prefixes. That explains why a bare /24 cannot be parsed; it does not prove which step inside a particular Hiddify build produced the empty field. Do not paste someone else's private key, full configuration, or "universal WARP parameters" over your own. ## Step 3 — Change one variable per attempt - Back up the current configuration, disable the built-in WARP first, and confirm the state of ordinary nodes. - If you use a custom warp://, check the address and fields you entered yourself, and rebuild only that WARP entry — leave the subscription and routing rules alone. - If you use the built-in generator, regenerate once through the current version's official interface and record whether the same error changes. - Check client versions only through official release channels. If the error survives an upgrade, keep the before/after versions, the platform and the reproduction steps for an issue report. Deleting app data also deletes the configuration and the diagnostic evidence, so it should not be step one. If you do need to reset, confirm the backup restores first, and follow the official instructions for your version. The public reports available do not show that a reset generally fixes this error. ## When to stop changing things yourself - The error involves a company or school VPN, a system network extension, or administrator rights. - You cannot confirm whether the configuration backup, WARP account details or subscription URL have been saved safely. - Regenerating several times changes nothing while the ordinary nodes still work. - The interface differs from this description and you cannot tell what the current control does. At that point, keep a minimal reproduction and a redacted log. If you need the ordinary import and connection path for Hiddify, work through a general Hiddify walkthrough. If you are comparing provider-side information, check published provider facts — and do not record a WARP parsing error as a provider outage. ## Sources Hiddify's official WARP guide for the built-in and custom flows; issues #930, #1389 and #1687 as public individual reports; the sing-box WireGuard address field documentation for the format requirement. This page did not reproduce those users' devices and has no access to their private WARP configuration. The recovery order is a diagnostic method built from public evidence, not a guarantee for every version. ## Frequently asked questions **Q: Does ParsePrefix("/24") mean the node is dead?** No. The error points at address-prefix parsing. Disable WARP and test the original node separately before drawing any conclusion about the node. **Q: Why does the error show only /24?** The text shows that the IP before the slash is empty — a syntax-level clue. Which field or which step produced the empty value has to be determined from the version and a redacted configuration. **Q: Will upgrading Hiddify definitely fix it?** There is no consistent evidence for that. The public issues span different platforms and versions and do not guarantee that an upgrade fixes your device. Keep the configuration and a reproduction record before upgrading. **Q: Can I just delete the app data directory?** Not as a first step. It removes both the configuration and the diagnostic evidence, and the public material does not show that it generally fixes this error. Back up, disable WARP, confirm ordinary connectivity, then follow the official recovery method for your version. **Q: Is this the same as having no nodes after a subscription update?** Different stage. This one happens while parsing the WARP outbound configuration. The other starts with subscription fetch, parsing and profile selection. ## Continue reading allowInsecure has been removed in v2rayNG · sing-box nftables: operation not supported · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/sing-box-auto-redirect-nftables-operation-not-supported/ TITLE: sing-box auto_redirect: nftables operation not supported LAST VERIFIED: 2026-11-14 sing-box · Linux TUN · nftables # sing-box auto_redirect: nftables operation not supported The tunnel interface came up, so the configuration was accepted. What failed is the step afterwards: the system refused one of the routing rules the core tried to install. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Split the failure by the first log line First key errorWhere it is stuckCheck nextpost-start inbound/tun ... auto-redirect: setup nftables: flush nftables followed by operation not supportedTUN was created, but the system rejected an automatic redirection ruleKeep the full log; check kernel, nftables, auto_route, auto_redirect and the rules your system actually acceptspermission denied, or missing permission to manage networkingPrivileges or container capabilitiesCheck the running identity, the container's network-management capability and how it was startednetwork: missing default interface with no fatal nftables error above itDefault egress or the network namespaceCheck the default route and interfaces; this warning alone is not a verdict on nftables supportLog points at route.rule_set download or parsingRemote rule set loadingUse a remote rule set checklist insteadCore started, ordinary HTTPS requests failConnection, DNS, routing or nodeVerify a real request and the egress ## What auto_redirect actually requires The official TUN configuration documentation states that auto_redirect is Linux-only, requires auto_route to be enabled at the same time, improves TUN routing through nftables, and needs kernel support for nfqueue for pre-matching. So the option depends on what the system's rule engine can do — but the documentation does not promise that every kernel and nftables combination supports it. ## One case that can be checked In sing-box issue #4588 the reporter used OpenWrt 23.05-SNAPSHOT on Linux 5.4.213, arm64, nftables 1.0.8, sing-box 1.14.2. The configuration created a TUN inside an isolated network namespace; with auto_route and auto_redirect enabled, the log first showed the TUN coming up and then reported operation not supported while installing the nftables rules. The reporter compared the same TCP reject rule at different hooks with nft -c and raised the question of kernel support ranges.The issue was still open at the time of checking, with no general fix published by a maintainer. It therefore cannot be used to claim that every Linux 5.4 device, every OpenWrt build or every sing-box version fails.The same log also contains network: missing default interface. That may relate to routing inside the isolated namespace and must not be treated on its own as the root cause of the auto_redirect failure. Judge from the earliest fatal chain, your current configuration, and what your own system supports. ## A recovery order that keeps an exit - Secure a way back. Work from a local terminal or a management path that does not depend on this proxy chain. Back up the sing-box configuration, the current network settings and the firewall rules. If remote management of the device depends on the same TUN, do not repeatedly restart or flush rules on a production router. - Record the environment. Save the output of sing-box version, uname -r and nft --version, and note whether you run on a host, in a container or on OpenWrt. Record whether both auto_route and auto_redirect are enabled, and whether nfnetlink_queue is loaded. A loaded module does not prove every nftables rule form is supported. - Separate static configuration from system application. Run sing-box check -c config.json on a copy to validate syntax and fields. Passing the check only means the static configuration parses — it does not mean the kernel accepts the nftables rules generated at startup. Then use the startup log to find the first FATAL, and do not count the repeated netlink errors after it as separate faults. - Compare under control. If you can preserve the original configuration and have a local way back, compare startup with auto_redirect on and off in an isolated environment. Turning it off changes how traffic is taken over, so "the process starts" must not be read as "all device traffic is proxied correctly". Do not swap the system's nftables tables on a production device. - Act on the evidence you have. If the error genuinely lands on a rule form the system does not support, check the firmware, kernel and sing-box version compatibility information first, then report the redacted log to the project or to the device maintainer. ## When to stop Stop testing if you cannot keep an exit path, if the device is the network gateway for the whole site, or if continuing requires flushing the production firewall. Those actions can take remote management and other devices offline. The namespace reproduction in the public issue describes that environment's error chain; it is not a configuration for everyone to copy, and the reporter's default_interface: lo is part of that isolated setup.This page did not reproduce the fault on the OpenWrt, Linux 5.4.213 and sing-box 1.14.2 environment described in #4588, and it has not measured that any particular kernel version always succeeds. If you are using Clash Verge Rev TUN or an ordinary system proxy, the configuration and takeover path differ — do not apply these Linux rule conclusions there. ## Sources Official TUN configuration field documentation; sing-box issue #4588 (OpenWrt 23.05-SNAPSHOT, Linux 5.4.213, nftables 1.0.8, sing-box 1.14.2, still open at the time of checking). Checked 2026-11-13. ## Frequently asked questions **Q: sing-box check passes, so why does it fail at runtime?** A static configuration check is not the same as the kernel accepting auto-generated nftables rules. Keep both the check output and the real startup log — they locate two different stages. **Q: Does 'missing default interface' mean Linux 5.4 is unsupported?** No. The isolated namespace in the public issue shows that warning together with the fatal nftables error, but the default-interface question needs its own look at the routing environment. Without the explicit auto-redirect: setup nftables chain, it is not the same fault. **Q: nfnetlink_queue is loaded but I still get 'operation not supported'. What now?** Identify the first rejected rule and your system version. The reporter in #4588 had the relevant module loaded and still failed on a particular TCP reject rule — a module being present does not guarantee every rule form is supported. **Q: Can I just turn auto_redirect off and keep using it?** Only as a controlled comparison, and only with the original configuration kept and a way to restore networking. The official documentation notes that this option changes Linux TUN routing; after disabling it you still have to verify device traffic, DNS and egress rather than judging by the service starting. **Q: Does this mean my provider's nodes are broken?** The error occurs while installing TUN rules on the local machine, which is too early to judge a remote node. Fix core startup and routing first, then make a real HTTPS request. ## Continue reading Hiddify WARP: netip.ParsePrefix with an empty IP · set system dns failed: exit code 1 · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/v2rayn-argumentoutofrange-count-subscription-empty/ TITLE: ArgumentOutOfRange_Count after a v2rayN update · Node Handbook LAST VERIFIED: 2026-11-14 v2rayN · subscription import · Windows # ArgumentOutOfRange_Count when updating subscriptions in v2rayN A browser opens the subscription link, so the fetch must be fine — except the error is not about the fetch. It appears after the response arrives, while v2rayN is turning that response into node records. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Put the error in the right stage first Log or symptomStageFirst actionTimeout, 403, TLS or DNS errorFetching the subscriptionCheck the access path, system clock, network and provider statusFetch succeeds, then ArgumentOutOfRange_CountParsing / importStop repeating the update; keep the old data and a redacted logUpdate finishes but the node count is 0Applying / replacingCheck the previous profile, a backup, or another deviceNodes exist but none of them connectCore, node, or system proxyMove to a connectivity and takeover checklistA browser opening the subscription proves only that a response was returned. It does not prove that v2rayN can identify the format. When the log already says the fetch succeeded, switching between direct and proxied access usually cannot explain the import exception that follows. ## Step 1 — Stop the operations that overwrite old data - Do not click "update all subscriptions" repeatedly. - Do not delete the subscription group, and do not run cleanup scripts of unknown origin. - Quit v2rayN normally and back up the current configuration directory. That backup may contain subscription tokens, so keep it somewhere trusted. - Record the v2rayN version, the OS version, the time of failure, and the redacted log starting from "fetch subscription content". - If another device still holds the old nodes, do not sync or update it — keep it as a reference you can restore from. ## Step 2 — Confirm what the subscription actually returns The v2rayN subscription documentation states that a subscription should be in a standard protocol format or an internal format v2rayN can read. Record only the HTTP status, the content type, the response size, whether it is empty, whether it is YAML, JSON, Base64 or HTML, and the protocol mix. Do not publish the full subscription.If the payload is Base64, decode it and check whether it is a list of standard URIs, one per line. Then determine whether one subscription fails, or every subscription fails on the same build.Do not paste the response into an online decoder, a forum post or a public issue. ## Step 3 — Narrow it down with a minimal comparison - Create a temporary subscription group that does not overwrite the original. - Use another compatible format that the provider officially offers; avoid public conversion sites. - Import a single small, redacted configuration of your own to see whether the parser accepts anything at all. - The same subscription working on an older build or another device proves only that the environments differ; it does not prove the new build is defective. - If several clients fail to read the same response, ask the provider to check the content first. ## Step 4 — Recover the node list safely - You have a directory backup: save the current failing directory, then restore the copy from before the failure. - You have a device that was not updated: keep an offline backup and do not export private nodes to a public location. - You only have the subscription endpoint: ask the provider to confirm the plan, token and return format, then test in a temporary group. - You have no backup: do not download unknown "repair database" tools. Keep the log and submit a minimal reproduction to the official project. Once the list is back, check the active profile before testing a single node. Do not enable TUN, change DNS and swap the core all at the same time. ## What to include in an issue report - The exact v2rayN version, the installation type, and the OS version. - Whether the error occurs at fetch, parse, import or apply. The node count before and after, plus response type, size and protocol mix.The first exception and its call stack, with subscription, IP, UUID, password, device name and private paths redacted.Whether the problem reproduces from an empty configuration directory or on another device. ## Sources and scope Material checked 2026-10-30 against v2rayN issue #10370, the official v2rayN subscription documentation, and v2rayN discussion #6699. Issues are public evidence from specific user environments; they do not mean the project has confirmed a general root cause, and they do not guarantee that upgrading or downgrading will fix your case.If the list is restored but nothing connects, the problem has moved to the connection and takeover stage. ## Three v2rayN failures that look like a core crash This page covers ArgumentOutOfRange_Count. Two other v2rayN errors end in the same “core failed” style dialog with a completely different cause. Identify which one your log shows before you touch the files. What the log or UI saysWhat it actually points to failed to open geosite.dat / ... > EOFGeo data file missing or truncated ArgumentOutOfRange_Count — this pagesubscription content empty, parser ran out of items allowInsecure has been removeddeprecated TLS field in an imported share link The dialog text is the same for all three; only the first kernel line differs. Quote that line, not the dialog title, when you ask for help or search for a fix. ## Frequently asked questions **Q: The browser opens the subscription, so why does v2rayN still fail?** A browser only retrieves the response. v2rayN still has to identify the format, parse each entry, and write local records. The failure can sit entirely in those later steps. **Q: Can I keep clicking update?** Not advisable. A failing path may replace or clear the existing records. Back up first and capture the first error instead. **Q: Can I post the subscription content in a GitHub issue?** No. Provide redacted format characteristics and a minimal reproduction configuration instead of the payload. **Q: Will downgrading v2rayN restore my nodes?** Not necessarily. Downgrading does not automatically restore data that has already been cleared, and the configuration structure may differ between versions. **Q: The list is back but nothing connects. What now?** Check the active server, whether the core starts, the system proxy or TUN state, and DNS. That is a separate stage from importing. **Q: How do I avoid losing nodes on the next update?** Keep a trusted backup, test format changes in a temporary group first, and keep one device that has not been updated as a reference. ## Continue reading failed to open geosite.dat in v2rayN · the remote profile data is invalid yaml · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/v2rayn-geosite-dat-missing/ TITLE: failed to open geosite.dat — v2rayN core won't start LAST VERIFIED: 2026-11-14 v2rayN · Windows / macOS / Linux # failed to open geosite.dat in v2rayN The alert v2rayN shows is a summary. The line that identifies the fault is the first error in the log above it — and a missing Geo data file looks very different from a truncated one, or from a port already in use. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Read the first core error, not the last one Failed to run Core is a result, not a cause. Open the v2rayN log and find the complete error chain that precedes it. Record the client version, the core you selected, the OS, the action that triggered it, and the file path quoted in the error. Never paste subscription URLs, node credentials or full configuration files into a public thread. ## Match the log fragment to the fault Log fragmentMore direct readingCheck firstfailed to open geosite.dat plus no such file or directoryThe path Xray looks in does not contain the fileThe exact path in the error, the selected core, and where the Geo files actually arefailed to check code ... from geosite.dat plus EOFThe file exists but is likely incomplete or malformedFile size, whether an update was interrupted, whether a regional preset was usedfailed to listen TCP plus address already in usePort conflict, not a Geo file faultA separate "core fails to start" checklistcan't download MMDB, HTTP 403/404The download step failedThe URL, HTTP status, current proxy and update settings — the file may never have been writtenIssue #9718 records a macOS log with bin/xray/geosite.dat: no such file or directory. In that specific packaging case a contributor suggested copying both geoip.dat and geosite.dat from bin into the xray directory, and later replies reported that copying alone still did not produce a connection. Treat it as one packaging-specific case, not a general fix, and do not copy it blindly on Windows or onto a different core directory without checking your own log. ## Recover in order - Back up the configuration. Use v2rayN's built-in backup, and note the regional preset currently selected. The official interface notes state that the backup covers the guiConfigs folder, and that on non-Windows builds the program directory and data directory may differ. - Locate the file from the log. Open the exact directory printed in the error. Check whether geosite.dat exists, whether its size is zero, and when it was last modified. Do not guess a fixed path from a blog post. - Check the selected core and the data source. Only if the selected Xray build actually needs the .dat files, re-fetch them through the Geo update entry that matches your v2rayN version. The update code builds separate requests for geosite.dat and geoip.dat, and the project wiki notes that the regional preset changes where Geo data and rule sets come from. If the update log shows 403, 404 or an interruption, fix access to the source before deciding the file was written. - Decide whether it is missing or corrupt. Missing: compare against the packaging case in #9718 for your core directory. EOF: stop restarting the core in a loop, fetch a complete file compatible with your current regional preset and core, and keep the old file as a backup. Do not make the message disappear by renaming the file, putting an empty file in place, or disabling validation. - Restart and read the new log. Only when the first Geo-file error is gone and the core starts normally has this stage been fixed. Then test the system proxy and a real page. A core that starts is not proof that your nodes work. ## Version and evidence boundary v2rayN's download layout and packaging change with version, OS and selected core. This page deliberately gives no fixed file-size threshold, and it does not recommend fetching Geo data from an unverified mirror. A public issue proves only that the reporter's scenario exists. Your own log path and HTTP status take precedence.If the symptom only appears after you enable TUN, the general permission and networking stages of TUN troubleshooting apply — but configuration steps are not transferable between different clients. ## Sources v2rayN issue #9718 (macOS, no such file or directory for the Xray Geo file), the v2rayN wiki and Geo update code for how geosite.dat and geoip.dat requests are generated, and the official interface notes on what the built-in backup covers. Checked 2026-11-05. ## Four Geo errors that look alike and are not v2rayN loads more than one data file at startup. geosite.dat carries domain rules, geoip.dat carries IP ranges, and cores that read the newer format use an .mmdb file instead. A failure in any of them produces a similar-looking line, so match the filename and the verb in your log before touching anything. Log lineWhat it actually meansFirst thing to check failed to open geosite.datThe domain list is not at the path the core searchedThe full path in the log, and which core is selected failed to open geoip.datThe IP list is missing, even when geosite.dat is presentWhether both files sit in the same directory ... from geosite.dat > EOFThe file opened but is incompleteFile size, and whether the last download finished can't download MMDBThe update request failed and nothing was writtenHTTP status, and whether the request went through a proxy ### failed to open geoip.dat The two .dat files are fetched and referenced as a pair, but nothing enforces that both land on disk. A download interrupted between the two requests leaves a working geosite.dat next to a missing geoip.dat, and the core only reports the second file after it has already read the first. Look at both filenames in the directory the log names, not only the one inside the error. ### can't download MMDB This line comes from the update path, not the startup path. It means the client tried to fetch the MMDB file and the request did not complete, so there is no file to open. Retrying the update while the client itself can reach the network is the normal fix; reinstalling the client is not, because the client was never the failing part. If the same log shows a 403 or 404, the source URL is the problem. ### The file exists and the core still refuses to start v2rayN can keep several cores side by side, each with its own directory. A .dat copied into one core's folder is invisible to another. Confirm the selected core first, then take the directory from the log line rather than from a guide — the packaged layout differs between the Windows build and the macOS build. ### The update says it succeeded but the file size did not change A Geo update that reports success while the file keeps its old timestamp usually means the write landed somewhere else, or the client is reading a cached copy. Record the modification time before and after, and check whether a regional preset is pointing the client at a different source. ## Three v2rayN failures that look like a core crash This page covers failed to open geosite.dat / ... > EOF. Two other v2rayN errors end in the same “core failed” style dialog with a completely different cause. Identify which one your log shows before you touch the files. What the log or UI saysWhat it actually points to failed to open geosite.dat / ... > EOF — this pageGeo data file missing or truncated ArgumentOutOfRange_Countsubscription content empty, parser ran out of items allowInsecure has been removeddeprecated TLS field in an imported share link The dialog text is the same for all three; only the first kernel line differs. Quote that line, not the dialog title, when you ask for help or search for a fix. ## Frequently asked questions **Q: Is 'Failed to run Core' always a missing geosite.dat?** No. It is a summary message. The steps on this page apply only when the log above it explicitly names a Geo file error. Port conflicts and invalid configuration fields have different fixes. **Q: Can I just download a geosite.dat and drop it in the program folder?** Not advised. Read the exact path from your log and identify the selected core first, then fetch the file through the update entry for your version. The program root is often not the directory the core reads from. **Q: What is the difference between 'no such file or directory' and EOF?** The first says the file is not where the core looked. The second says reading the data ended early, which points at an interrupted download, a bad payload, or a preset mismatch. Re-importing the subscription fixes neither. **Q: The Geo update says it finished, but the error remains. Why?** Check the target path in the log, the file modification time, and which core is selected. A finished download does not prove the selected core reads that directory, and an HTTP error during the update means the data may be incomplete. **Q: After the fix, node latency still shows -1 ms. Did the Geo fix fail?** Not necessarily. First confirm the new log has no Geo file error, then work through core startup and node connectivity as separate stages. A short latency reading on its own does not prove node quality. ## Continue reading allowInsecure has been removed in v2rayNG · ArgumentOutOfRange_Count when updating subscriptions · Full error string index ---------------------------------------------------------------------- URL: https://jiedianshouce.com/en/articles/v2rayng-allowinsecure-removed-tls-config/ TITLE: allowInsecure has been removed — v2rayNG TLS failure LAST VERIFIED: 2026-11-14 v2rayNG · Android · Xray core # Failed to build TLS config: allowInsecure has been removed in v2rayNG The node list shows -1 ms for every TLS node and the log refuses to load the configuration. The error is a format rejection, not a network timeout — and the usual advice to turn off certificate checks is the wrong direction. Node Handbook editorial team · verified 2026-11-14Material checked on 2026-11-14. Error strings are quoted verbatim so you can match them against your own log. ## Judge by the exact error, not by the symptom Log or symptomWhich step to judgeDo not do thisFailed to build TLS config immediately followed by allowInsecure has been removedAn old TLS field is incompatible with the current core's configuration formatDo not keep re-testing latency, and do not declare every node deadConfiguration loads, then x509, a certificate name mismatch, or a TLS handshake failureThe certificate or handshake stage — check SNI, the certificate and the device clockDo not file it as the removed-field parsing errorOnly dial tcp ... timeoutA connection attempt was made; check address, port and network pathDo not touch TLS toggles to make a timeout disappearNode connects, but specific sites failDNS, routing and egressDo not rewrite every node in the subscriptionIssue #6141 in the v2rayNG repository records one case: Android 16, v2rayNG 2.3.5 from F-Droid, Xray-core 26.7.28. Several TLS nodes reported -1 ms and the log rejected allowInsecure while loading the configuration. A maintainer closed the issue and pointed at the related explanation. That is a specific environment report — it does not prove every v2rayNG version and every TLS node behaves the same way. ## Safe handling order - Keep minimal evidence. Note the v2rayNG app version, the core version, the node protocol, the exact error text and when it happened. Redact the subscription URL, password, UUID and server address before sharing a screenshot. Keep the original subscription and a copy of the node configuration. - Decide whether the configuration loaded at all. Open the log for a single node and look for Failed to build TLS config together with allowInsecure. If those two lines are absent, go back to the table above; the steps here do not apply. - Check where the configuration came from. If the node arrived through an automatic subscription update, check whether the source still emits allowInsecure, insecure=1, or a "skip certificate verification" flag. Locate the field only in your own copy, and do not send a private subscription to a public conversion site. - Ask the source for a compatible configuration. For a paid service, the provider should check the certificate, SNI, subscription output and the fields the current client supports. If you run the server yourself, fix the server-side certificate and domain configuration. Changing allowInsecure=true to false and declaring victory is not a fix: if the node depended on skipping verification, you will simply meet a real certificate error next. - Verify client and core versions from the official channel. The v2rayNG Releases page carries version and compatibility notes. Back up the configuration before upgrading or downgrading. An older build that still accepts insecure fields is not a security fix. - Re-verify. After updating the configuration, confirm the original error is gone and the node entry still exists, then test a single connection, an ordinary HTTPS page, and your egress IP. If the error becomes x509 or a timeout, keep diagnosing from that new error — the old message disappearing is not the same as the whole chain working. ## Why you should not simply disable certificate checks allowInsecure concerns verification of the proxy server's identity. Skipping the certificate check lowers the strength of that verification, and public nodes in particular are unsuitable for sensitive accounts. The v2rayNG release notes point towards certificate pinning or a server-side change instead.A certificate fingerprint has to come from a channel you can independently verify; do not copy one out of the same untrusted connection. If you cannot verify the server certificate or the provider's identity, stopping use of that node is a better outcome than switching verification off. ## Scope and evidence boundary This page is built from an official user report and the project's release notes. It was not reproduced across every Android version, core, subscription format and carrier. The detail in #6141 that the same subscription worked on a desktop client describes the reporter's comparison only; it does not prove Android is the sole failing point. If the log only contains the words allowInsecure without surrounding context, record the complete error chain first. TLS and REALITY parameters are not interchangeable between protocols.Checked 2026-11-12 against v2rayNG issue #6141, the v2rayNG Releases page, and the maintainer-linked discussion on migrating away from allowInsecure. ## Four TLS failures that arrive through the same dialog v2rayNG reports almost every TLS problem as Failed to build TLS config, and beginners read all of them as "the certificate option is gone". Only one of the four below is about that option. The text after the colon is what separates them. Error textWhat failedWhere to look allowInsecure has been removedThe config still carries a field the current core no longer acceptsThe node's TLS settings, and whether it was imported from an old share link x509: certificate signed by unknown authorityHandshake reached the server, but the chain could not be verifiedWhether the server uses a self-signed or privately issued certificate x509: certificate has expired or is not yet validThe certificate is outside its validity windowThe device clock, then the server certificate dates EOF during the handshakeNothing usable came back at allPort, SNI and whether the server accepts TLS on that port ### x509: certificate signed by unknown authority This is a verification failure, not a configuration syntax error. The client completed the TCP and TLS hello, then rejected the chain. It usually means the server presents a self-signed certificate, or the device is behind a network that terminates TLS. The fix belongs on the server side or in a proper pinned certificate, not in disabling verification for every node you own. ### x509: certificate has expired or is not yet valid Before you ask the provider anything, check the device clock. A phone that is minutes or hours off will reject a perfectly valid certificate, and the error text is identical to a genuinely expired one. If the clock is right, the certificate dates are the server operator's problem. ### EOF during the handshake EOF means the connection closed before a TLS record came back. That points at the port or the SNI, not at certificates: a TLS port that expects a different hostname, a server that closed the connection because the path was wrong, or a middlebox resetting the flow. It cannot be fixed by relaxing certificate checks, because no certificate was ever received. ### The node works in another client but not in v2rayNG Compare what each client actually sends: transport, path, host header and SNI. v2rayNG drops fields it no longer supports when it imports an old link, so a node that worked in an older version can import "successfully" and still fail at the TLS stage. That is an import-fidelity problem, and the tell is a config field that silently disappeared after import. ## Three v2rayN failures that look like a core crash This page covers allowInsecure has been removed. Two other v2rayN errors end in the same “core failed” style dialog with a completely different cause. Identify which one your log shows before you touch the files. What the log or UI saysWhat it actually points to failed to open geosite.dat / ... > EOFGeo data file missing or truncated ArgumentOutOfRange_Countsubscription content empty, parser ran out of items allowInsecure has been removed — this pagedeprecated TLS field in an imported share link The dialog text is the same for all three; only the first kernel line differs. Quote that line, not the dialog title, when you ask for help or search for a fix. ## Frequently asked questions **Q: All nodes show -1 ms. Is that this error?** No. -1 is the result of a failed test and can come from configuration parsing, networking, certificates, DNS or the test method itself. Only when the log explicitly contains Failed to build TLS config and allowInsecure has been removed should you use this page. **Q: If I delete the allowInsecure field, will it connect?** Not necessarily. Removing the old field may let the configuration reach the next stage; if the server certificate cannot be validated, a certificate error follows. The provider should supply a trusted certificate or a configuration you can verify, and you should still make a real HTTPS request. **Q: Can I just install an older v2rayNG?** Only as a short-term comparison. Treating acceptance of an insecure field as a fix is wrong. Get builds from official Releases and back up your configuration; do not install modified builds from unknown sources. **Q: The error returns after a subscription update. Why?** The source is probably still emitting the old field, or the client overwrote your local change during the update. Keep a redacted error and ask the provider to change the subscription output instead of editing your only copy again and again. **Q: The same subscription works on my computer but not on my phone. Does that prove the nodes are fine?** It only proves they worked on that computer with that client, core and network at that moment. On the phone you still need the app version, the generated configuration and the full log before drawing a conclusion. ## Continue reading failed to open geosite.dat in v2rayN · Hiddify WARP: netip.ParsePrefix with an empty IP · Full error string index