[GOOD HELP]Reload needed many times to work

I have quite a few subdomains, each with proxied CNAME records. Now when I create a new subdomain, and you try to visit any page on said subdomain, you have to reload the page a lot of times before it shows up. Here’s how stuff gets from my servers to Cloudflare: Server->cloudflared->Cloudflare On valid pages, it will still show up as blank(404 in dev console) until you reload it several times. I have no idea why this happens, nor how to fix it. I have tried deleting unused DNS records. Edit: To fix this on Linux, run the following command, then start cloudflared again: sudo pkill -f cloudflared It will kill all instances of the tunnel, and you can then start cloudflared again, and it will have the new configuration.
5 Replies
Alexm
AlexmOP2w ago
Note: Caching should not be an issue, as every URL for this particular thing is random. Also, caching for this is disabled.
Idle
Idle2w ago
sounds like a caching issue
Alexm
AlexmOP2w ago
I don’t have caching enabled for this. Cache rule 1:
(http.request.uri.path.extension eq "mp3") or (http.request.uri.path.extension eq "png") or (http.request.uri.path.extension eq "jpg") or (http.request.uri.path.extension eq "jpeg") or (http.request.uri.path.extension eq "mov") or (http.request.uri.path.extension eq "gif") or (http.request.uri.path.extension eq "avi") or (http.request.uri.path.extension eq "flv") or (http.request.uri.path.extension eq "bmp") or (http.request.uri.path.extension eq "tiff") or (http.request.uri.path.extension eq "wav") or (http.request.uri.path.extension eq "ogg") or (http.request.uri.path.extension eq "zip") or (http.request.uri.path.extension eq "pdf") or (http.request.uri.path.extension eq "apk") or (http.request.full_uri wildcard "https://creators.atdevs.org/*")
(http.request.uri.path.extension eq "mp3") or (http.request.uri.path.extension eq "png") or (http.request.uri.path.extension eq "jpg") or (http.request.uri.path.extension eq "jpeg") or (http.request.uri.path.extension eq "mov") or (http.request.uri.path.extension eq "gif") or (http.request.uri.path.extension eq "avi") or (http.request.uri.path.extension eq "flv") or (http.request.uri.path.extension eq "bmp") or (http.request.uri.path.extension eq "tiff") or (http.request.uri.path.extension eq "wav") or (http.request.uri.path.extension eq "ogg") or (http.request.uri.path.extension eq "zip") or (http.request.uri.path.extension eq "pdf") or (http.request.uri.path.extension eq "apk") or (http.request.full_uri wildcard "https://creators.atdevs.org/*")
Any of those are marked as eligible for caching. At first I thought that maybe moving from one server(that has a ton of things on one tunnel) to another(that has barely anything on its tunnel) would work, but it didn’t. Bad response(no content): Cache response: DYNAMIC Response code: 404 Good response(content shows): Cache response: DYNAMIC Response code: 200 Caching should not be an issue. Each URL is randomly generated. Each URL is only used once. Cloudflare seems to be broken
The_Cookie
The_Cookie7d ago
Maybe try to enable that development mode which temporarily disables any caching feature
Alexm
AlexmOP7d ago
The origin never sees the request. I get a 404, but the output of the test Python Flask server never shows any requests. Request route: Request->Cloudflare->Tunnel->Web server The web server seems fine, as it’s accessible over Tailscale, the tunnel works, as everything else other than new things works. Nope. Locally. (~/.cloudflared/config.yml) The weird thing is that https://news.atdevs.org works, but everything else doesn’t. I have SIGKILLed all cloudflared processes before. Doesn’t work right. on either of my two servers. I’ll re-kill cloudflared. Different tunnels. Same configs wouldn’t work. It would appear that my custom server management software had a stroke and died. I told it to kill cloudflared, but it never did. Connectors existed with different configs loaded in RAM. sudo pkill -f cloudflared and sudo pkill -f atav-mon-restart-on-crash fixed it. I meant good help as in it’s useful for people having the same issue.

Did you find this page helpful?