Nginx failing to resolve upstream with custom DNS resolver - nginx

docker run --rm --net=host -v $PWD/default.conf:/etc/nginx/conf.d/default.conf nginx
2019/05/12 17:02:49 [emerg] 1#1: host not found in upstream "tickethub.service.consul" in /etc/nginx/conf.d/default.conf:10
nginx: [emerg] host not found in upstream "tickethub.service.consul" in /etc/nginx/conf.d/default.conf:10
While dig shows the DNS record correctly:
dig #127.0.0.1 -p 8600 tickethub.service.consul
; <<>> DiG 9.12.3-P1 <<>> #127.0.0.1 -p 8600 tickethub.service.consul
; (1 server found)
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 57394
;; flags: qr aa rd; QUERY: 1, ANSWER: 3, AUTHORITY: 0, ADDITIONAL: 4
;; WARNING: recursion requested but not available
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 4096
;; QUESTION SECTION:
;tickethub.service.consul. IN A
;; ANSWER SECTION:
tickethub.service.consul. 0 IN A 172.23.0.6
tickethub.service.consul. 0 IN A 172.23.0.5
tickethub.service.consul. 0 IN A 172.23.0.7
;; ADDITIONAL SECTION:
tickethub.service.consul. 0 IN TXT "consul-network-segment="
tickethub.service.consul. 0 IN TXT "consul-network-segment="
tickethub.service.consul. 0 IN TXT "consul-network-segment="
;; Query time: 0 msec
;; SERVER: 127.0.0.1#8600(127.0.0.1)
;; WHEN: Sun May 12 16:58:54 GMT 2019
;; MSG SIZE rcvd: 209
And my nginx config:
server {
listen 80;
server_name localhost;
location / {
resolver 127.0.0.1:8600;
proxy_pass http://tickethub.service.consul;
}
error_page 500 502 503 504 /50x.html;
location = /50x.html {
root /usr/share/nginx/html;
}
}
What may be the issue?

This worked when I explicitly set the DNS on the docker container to 127.0.0.1 which means Nginx is probably trying to resolve it WITHOUT using the resolver specified argh...
I think I also had to change the DNS port to 53 instead of the explicit 8600
Or something...
Probably a bunch of nginx bugs...

mic drop
It worked when I set the proxy_pass using a variable:
location / {
resolver consul;
set $endpoint tickethub.service.consul;
proxy_pass http://$endpoint/;
}

Related

How to make grafana on nixos available in local network

My laptop and my nixos-server (hostname=nixos) are both conected to my router (fritz.box). I can access the rooter via ping (ping nixos.fritz.box) and ssh (ssh username#nixos.fritz.box).
What I want is to follow the first part of this guide to set up grafana on nixos. I then want to be able to access grafana from my laptop.
On the server I have configured nixos to run both grafana and a reverse proxy (nginx):
services.grafana = {
enable = true;
domain = "grafana.nixos.fritz.box";
port = 2342;
addr = "127.0.0.1";
};
# nginx reverse proxy for grafana
services.nginx.virtualHosts.${config.services.grafana.domain} = {
locations."/" = {
proxyPass = "http://127.0.0.1:${toString config.services.grafana.port}";
proxyWebsockets = true;
};
};
# Open ports for http and https
networking.firewall.allowedTCPPorts = [ 80 443 ];
system.stateVersion = "21.03";
Unfortunatelly I can't access the grafana webinterface from my laptop.
I tried changing around the value of services.grafana.domain and what I type into my browser (firefox/curl), here is what I got:
services.grafana.domain
argument of curl
output of curl
grafana.nixos.fritz.box
http://grafana.nixos.fritz.box/
curl: (6) Could not resolve host: grafana.nixos.fritz.box
grafana.nixos.fritz.box
https://grafana.nixos.fritz.box/
curl: (6) Could not resolve host: grafana.nixos.fritz.box
grafana.nixos.fritz.box
http://nixos.fritz.box/
curl: (52) Empty reply from server
grafana.nixos.fritz.box
https://nixos.fritz.box/
curl: (35) LibreSSL SSL_connect: SSL_ERROR_SYSCALL in connection to nixos.fritz.box:443
nixos.fritz.box
http://nixos.fritz.box/
curl: (52) Empty reply from server
nixos.fritz.box
https://nixos.fritz.box/
curl: (35) LibreSSL SSL_connect: SSL_ERROR_SYSCALL in connection to nixos.fritz.box:443
grafana.localhost
(on the server) http://grafana.localhost
curl: (7) Failed to connect to grafana.localhost port 80: Connection refused
grafana.localhost
(on the server) https://grafana.localhost
curl: (7) Failed to connect to grafana.localhost port 443: Connection refused
Especially the last 2 lines leave me perplexed.
netstat -an | grep LISTEN on the server gives me this:
tcp 0 0 0.0.0.0:22 0.0.0.0:* LISTEN
tcp 0 0 127.0.0.1:2342 0.0.0.0:* LISTEN
tcp6 0 0 :::22 :::* LISTEN
unix 2 [ ACC ] STREAM LISTENING 1837 /run/systemd/private
unix 2 [ ACC ] STREAM LISTENING 1841 /run/systemd/userdb/io.systemd.DynamicUser
unix 2 [ ACC ] SEQPACKET LISTENING 1853 /run/systemd/coredump
unix 2 [ ACC ] STREAM LISTENING 1862 /run/systemd/journal/stdout
unix 2 [ ACC ] SEQPACKET LISTENING 1868 /run/udev/control
unix 2 [ ACC ] STREAM LISTENING 26958 /var/run/nscd/socket
unix 2 [ ACC ] STREAM LISTENING 1905 /run/systemd/journal/io.systemd.journal
unix 2 [ ACC ] STREAM LISTENING 12193659 /run/user/1001/bus
unix 2 [ ACC ] STREAM LISTENING 12205464 /run/user/1001/systemd/private
unix 2 [ ACC ] STREAM LISTENING 13312 /nix/var/nix/daemon-socket/socket
unix 2 [ ACC ] STREAM LISTENING 18416 /var/run/dhcpcd.sock
unix 2 [ ACC ] STREAM LISTENING 18418 /var/run/dhcpcd.unpriv.sock
unix 2 [ ACC ] STREAM LISTENING 13308 /run/dbus/system_bus_socket
I don't know how to make grafana available in the local network. Can someone help me with that, please?
(I know this question is somewhat similar to this one, but the solution there doesn't help me)
Adding the following line solved my problem (thanks to #Tch):
services.nginx.enable = true;

How to point custom domain to VM instance

These are my dns records:
Name | Type | TTL | Target
A 3600 185.199.108.153
A 3600 185.199.109.153
A 3600 185.199.110.153
A 3600 185.199.111.153
www CNAME 3600 pushp1997.github.io
A 3600 34.71.130.252
Here, the first 5 entries are to open my static github pages site.
The last entry is the ip address of my VM instance on GCloud.
These are my nginx server settings:
server {
listen 80;
server_name pushp.ml;
location /linkedin {
return 302 https://in.linkedin.com/in/pushp-vashisht;
}
}
Now, if I try 34.71.130.252/linkedin it redirects me to https://in.linkedin.com/in/pushp-vashisht.
But, when I try pushp.ml/linkedin it shows a 404 page of Github Pages.
How do I make pushp.ml/linkedin to redirect to https://in.linkedin.com/in/pushp-vashisht?
Edit:
On running dig command:
$ dig pushp.ml
; <<>> DiG 9.11.3-1ubuntu1.12-Ubuntu <<>> pushp.ml
;; global options: +cmd
;; Got answer:
;; ->>HEADER<<- opcode: QUERY, status: NOERROR, id: 36898
;; flags: qr rd ra; QUERY: 1, ANSWER: 5, AUTHORITY: 0, ADDITIONAL: 1
;; OPT PSEUDOSECTION:
; EDNS: version: 0, flags:; udp: 65494
;; QUESTION SECTION:
;pushp.ml. IN A
;; ANSWER SECTION:
pushp.ml. 3600 IN A 185.199.109.153
pushp.ml. 3600 IN A 185.199.110.153
pushp.ml. 3600 IN A 34.71.130.252
pushp.ml. 3600 IN A 185.199.108.153
pushp.ml. 3600 IN A 185.199.111.153
;; Query time: 431 msec
;; SERVER: 127.0.0.53#53(127.0.0.53)
;; WHEN: Fri Jul 17 17:13:57 EDT 2020
;; MSG SIZE rcvd: 117
In answer section we can see that ip addresses of github pages server as well as my vm instance are there.

What request path is location = /bar supposed to match in Nginx?

What request path is location = /bar supposed to match in Nginx?
What works fine: location /bar
Here is my nginx configuration, host file configuration, and my HTML files.
# cat /etc/nginx/sites-enabled/foo
server {
listen 80;
listen [::]:80;
server_name foo;
root /tmp/;
location /bar/ {
alias /var/www/foo/;
}
}
# cat /etc/hosts
127.0.0.1 localhost foo
127.0.1.1 debian
# cat /tmp/index.html
Hi! I am Tmp!
# cat /var/www/foo/index.html
<p>Hi! I am Index!</p>
# cat /var/www/foo/max.html
<p>Hi! I am Max!</p>
HTTP requests to the root, /bar/, and /bar/max.html produce the expected
output:
# systemctl restart nginx && curl http://foo/
Hi! I am Tmp!
# systemctl restart nginx && curl http://foo/bar/
<p>Hi! I am Index!</p>
# systemctl restart nginx && curl http://foo/bar/max.html
<p>Hi! I am Max!</p>
What does not work fine: location = /bar
Now I edit the configuration to replace location /bar with location = /bar:
# cat /etc/nginx/sites-enabled/foo
server {
listen 80;
listen [::]:80;
server_name foo;
root /tmp/;
location = /bar/ {
alias /var/www/foo/;
}
}
# systemctl restart nginx && curl http://foo/
Hi! I am Tmp!
These HTTP requests no longer work:
# systemctl restart nginx && curl http://foo/bar/
<html>
<head><title>404 Not Found</title></head>
<body bgcolor="white">
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.10.3</center>
</body>
</html>
# tail -n 1 /var/log/nginx/error.log
2018/05/25 00:07:50 [error] 29157#29157: *1 open() "/tmp/bar/index.html" failed (2: No such file or directory), client: 127.0.0.1, server: foo, request: "GET /bar/ HTTP/1.1", host: "foo"
# systemctl restart nginx && curl http://foo/bar
<html>
<head><title>404 Not Found</title></head>
<body bgcolor="white">
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.10.3</center>
</body>
</html>
# tail -n 1 /var/log/nginx/error.log
2018/05/25 00:08:49 [error] 29203#29203: *1 open() "/tmp/bar" failed (2: No such file or directory), client: 127.0.0.1, server: foo, request: "GET /bar HTTP/1.1", host: "foo"
# systemctl restart nginx && curl http://foo/bar/max.html
<html>
<head><title>404 Not Found</title></head>
<body bgcolor="white">
<center><h1>404 Not Found</h1></center>
<hr><center>nginx/1.10.3</center>
</body>
</html>
# tail -n 1 /var/log/nginx/error.log
2018/05/25 00:10:10 [error] 29265#29265: *1 open() "/tmp/bar/max.html" failed (2: No such file or directory), client: 127.0.0.1, server: foo, request: "GET /bar/max.html HTTP/1.1", host: "foo"
It appears that the GET requests for either /bar or /bar/ match the location = /bar/ directive? I thought these requests should have worked with this directive because http://nginx.org/en/docs/http/ngx_http_core_module.html#location mentions:
Also, using the “=” modifier it is possible to define an exact match of URI and location. If an exact match is found, the search terminates.
But as explained in my example, it does not seem to work? What kind of request would match the location = /bar directive then?
The URI /bar/ relies on the index directive to internally rewrite the URI to /bar/index.html. See this document for details.
The exact match location syntax will only match the original URI, and not the rewritten URI.
nginx will process the rewritten URI using the default location (which in your configuration, is the server context). So, the URI /bar/index.html will be searched for at /tmp/bar/index.html, and not found.

How to use The uWSGI FastRouter whith Nginx?

Configuration of Nginx:
location / {
include uwsgi_params;
uwsgi_pass 127.0.0.1:3030;
}
uWSGI FasterRouter
uwsgi --fastrouter 127.0.0.1:3030 --fastrouter-subscription-server 127.0.0.1:3131 -M
uWSGI web-app instance
uwsgi -M --subscribe-to 127.0.0.1:3131:/ --file server.py --http :8080
then HTTP GET [server ip]:80/ no response
HTTP GET [server ip]:8080/ got response
Configuration of Nginx:
location / {
include uwsgi_params;
uwsgi_pass 127.0.0.1:3030;
}
uwsgi -M --file server.py -s 127.0.0.1:3030
then HTTP GET [server ip]:80/ Got response
Do as follows. Worked. Thanks roberto for help.
Configuration of Nginx:
location / {
include uwsgi_params;
uwsgi_pass 127.0.0.1:3030;
}
uWSGI FasterRouter
uwsgi --fastrouter 127.0.0.1:3030 --fastrouter-subscription-server 127.0.0.1:3131 -M
uWSGI web-app instance
uwsgi -M --subscribe-to 127.0.0.1:3131:[server_ip] --file server.py -s 127.0.0.1:3232
Console output
[uwsgi-subscription for pid 18957] new pool: [server_ip] (hash key: 22902)
[uwsgi-subscription for pid 18957] [server_ip] => new node: 127.0.0.1:3232
then HTTP GET [server ip]:80/ GOT response

NGINX not following alias directive

I am trying to get nginx to server some static files. I put in what I believe should be the proper directive to alias the url, but nginx is refusing to server the page. My server.conf is as follows:
server {
listen 80 default_server;
listen [::]:80 default_server ipv6only=on;
root /usr/share/nginx/html;
index index.html index.htm index.php;
# Make site accessible from server.fqdn.suffix
server_name server.fqdn.suffix;
location /static/ {
autoindex on;
alias /usr/share/nginx/html/poindexter/inspire/static/;
}
# Include the basic h5bp config set
include /etc/nginx/h5bp/basic.conf;
}
I assumed that this would redirect to the /usr/share/nginx/html/poindexter/inspire/static/ directory. And for the autoindex it seems to work. But when you click on a file it generates a 404 error. I took an strace and sure enough, it is not honoring the alias.
[pid 2542] <... epoll_wait resumed> {{EPOLLIN, {u32=1822502928, u64=139901792235536}}}, 512, 4294967295) = 1
[pid 2542] accept4(10, {sa_family=AF_INET, sin_port=htons(37845), sin_addr=inet_addr("198.55.232.86")}, [16], SOCK_NONBLOCK) = 24
[pid 2542] epoll_ctl(21, EPOLL_CTL_ADD, 24, {EPOLLIN|EPOLLET, {u32=1822503697, u64=139901792236305}}) = 0
[pid 2542] epoll_wait(21, {{EPOLLIN, {u32=1822503697, u64=139901792236305}}}, 512, 10000) = 1
[pid 2542] recvfrom(24, "GET /static/css/bootstrap.min.cs"..., 1024, 0, NULL, NULL) = 772
[pid 2542] open("/usr/share/nginx/html/static/css/bootstrap.min.css", O_RDONLY|O_NONBLOCK) = -1 ENOENT (No such file or directory)
[pid 2542] write(8, "2014/03/05 21:59:23 [error] 2542"..., 328) = 328
[pid 2542] writev(24, [{"HTTP/1.1 404 Not Found\r\nServer: "..., 172}, {"<html>\r\n<head><title>404 Not Fou"..., 116}, {"<hr><center>nginx</center>\r\n</bo"..., 46}], 3) = 334
[pid 2542] setsockopt(24, SOL_TCP, TCP_NODELAY, [1], 4) = 0
[pid 2542] recvfrom(24, 0x14e9240, 1024, 0, 0, 0) = -1 EAGAIN (Resource temporarily unavailable
I would really appreciate any help I could get. I am very stuck.
I see several way to fix this.
1
Just add symlink from /usr/share/nginx/html/static/ to /usr/share/nginx/html/poindexter/inspire/static/ and remove alias directive. It something like aliasing on file system level.
location /static/ {
autoindex on;
}
2
Remove location ~* \.(?:css|js)$ { block.
3
Use rewrite (this solution make use the fact you static directory in inside your root)
location ^~ /static/ {
autoindex on;
rewrite ^(.+)$ /poindexter/inspire$1;
}

Resources