Issue upstream timed out (110: Connection timed out) while reading response header from upstream

[varoon]

Server operating system version

Ubuntu 20.04.5 LTS

Plesk version and microupdate number

Plesk Obsidian 18.0.47

Hi,

Im new to plesk and usgin it on linode.

We have a website developed in ReactJS, NodeJS and mongo db.

1849#0: *327290 upstream timed out (110: Connection timed out) while reading response header from upstream

Attached is the log on Nginx error.



My Nginx config file: /etc/nginx/nginx.conf

#user  nginx;
worker_processes  1;

#error_log  /var/log/nginx/error.log;
#error_log  /var/log/nginx/error.log  notice;
#error_log  /var/log/nginx/error.log  info;

#pid        /var/run/nginx.pid;

include /etc/nginx/modules.conf.d/*.conf;

events {
    worker_connections  2048;
}


http {
    include       mime.types;
    default_type  application/octet-stream;

    #log_format  main  '$remote_addr - $remote_user [$time_local] "$request" '
    #                  '$status $body_bytes_sent "$http_referer" '
    #                  '"$http_user_agent" "$http_x_forwarded_for"';

    #access_log  /var/log/nginx/access.log  main;

    sendfile        on;
    #tcp_nopush     on;

    #keepalive_timeout  0;
    keepalive_timeout  65;
    #tcp_nodelay        on;

    #gzip  on;
    #gzip_disable "MSIE [1-6]\.(?!.*SV1)";

    server_tokens off;

    include /etc/nginx/conf.d/*.conf;

# added to test
        proxy_http_version 1.1;
        proxy_set_header Connection "";
}

# override global parameters e.g. worker_rlimit_nofile
include /etc/nginx/*global_params;

what is the reason for this error?

Any help on how to get rid of these?

 

[IgorG]

Perhaps it is caused by the default timeout limit is 60 seconds for proxying requests from nginx to Apache. But the script takes more than 60 seconds to execute.
Try to add in Additional nginx directives field of domain Nginx settings to increase the timeout limit to 180 seconds (3 minutes):

proxy_connect_timeout 180s;
proxy_send_timeout 180s;
proxy_read_timeout 180s;
fastcgi_send_timeout 180s;
fastcgi_read_timeout 180s;

 

[varoon]

Thanks for the reply.

ok, i will add them inside the http right?

http {
......

proxy_connect_timeout 180s;
proxy_send_timeout 180s;
proxy_read_timeout 180s;
fastcgi_send_timeout 180s;
fastcgi_read_timeout 180s;

......
}

Will test and update...

 

[varoon]

just a query:
Since we use ReactJS, NodeJS and MongoDB web app, Do I need Apache?
can I disable Apache? will it affect the website?

 

[Yaroslav_T]

Hi!
In the current implementation, you can't disable Apache at all on the server level. However, you can do it for a single domain or service plan: https://support.plesk.com/hc/en-us/articles/213924385

The timeouts above are to be added in Plesk > Domains > example.com > Apache & nginx settings > Additional nginx directives. See this for reference: https://support.plesk.com/hc/en-us/articles/115000170354

In general, "110 timeout error" is very general and means that nginx waited for some time (60 sec, by default) to get the answer from Apache (upstream) but got nothing in return. SO increasing timeout will either fix the issue at all or will probably give a clue of what is happening

 

[varoon]

Perhaps it is caused by the default timeout limit is 60 seconds for proxying requests from nginx to Apache. But the script takes more than 60 seconds to execute.
Try to add in Additional nginx directives field of domain Nginx settings to increase the timeout limit to 180 seconds (3 minutes):

proxy_connect_timeout 180s;
proxy_send_timeout 180s;
proxy_read_timeout 180s;
fastcgi_send_timeout 180s;
fastcgi_read_timeout 180s;

After adding this to the http block, i restarted the Linode serve, the website did not work!
reverting the original conf file, the server started.

Actually, yesterday I added this mentioned from some other blog, (proxy_connect_timeout 180s; and other values)
created a file and pasted on this location: /etc/nginx/conf.d/timeout.conf

Am i doing anything wrong? Did i miss someting?

 

[Bitpalast]

After you added the timeout.conf file: What is the result of
# nginx -t
?

 

[varoon]

After you added the timeout.conf file: What is the result of
# nginx -t
?

it throws duplicate proxy_connect_timeout
So removed and restarted Linode. server

Now I receive a slightly different errors

Totally 3 different:

#1: No more recent errors (Fixed I think, let me check and update if it occurs again)

1849#0: *327290 upstream timed out (110: Connection timed out) while reading response header from upstream

#2: still occurs

2091#0: *29828 connect() failed (110: Connection timed out) while connecting to upstream

#3: still occurs

3231#0: *13332 peer closed connection in SSL handshake (104: Connection reset by peer) while SSL handshaking to upstream

 

[IgorG]

Am i doing anything wrong? Did i miss someting?

Why do you edit the nginx global configuration file manually instead of making these changes through the Plesk UI for a specific domain, as I wrote above?

 

[Bitpalast]

and restarted Linode. server

Instead of restarting the whole server, have you also tried to restart Nginx and Apache? Like
# service apache2 restart
# service nginx restart
?

 

[varoon]

Why do you edit the nginx global configuration file manually instead of making these changes through the Plesk UI for a specific domain, as I wrote above?

My Bad!. You are right. Actually, i did not know we can do on the Plesk itself. did not understand that.
Now Added them as mentioned Thanks!

 

[varoon]

Instead of restarting the whole server, have you also tried to restart Nginx and Apache? Like
# service apache2 restart
# service nginx restart
?

thanks, I will try that

 

[varoon]

Does not seem to solve the issue.

What else should I do to fix this?

I still get errors

2070#0: *16414 upstream timed out (110: Connection timed out) while reading response header from upstream

 

[varoon]

How do I check if the proxy_connect_timeout 180s; is applied?