Bad request 400

[leonep]

Hi,
i have occasonaly errors on various website in my server ( nginx reverse proxy ea-nginx, apache 2.4 )
Also i get errors from google page indexing issues 400.
i try to flush dns cache on clients, i try to found somwthing on logs but i found nothing related.
i try to set nginx error log on info but nothing appear.
may be i need debug to find this errors ?

before activate debug log level globally i try to set debug on a site with this:
Efficiently Troubleshooting NGINX with Custom Error Log (error_log) Configuration Options

but after restart nginx i got Forbidden on the site..
there is particulare permissions settings to setup ?

thanks

 

[cPRex]

Hey hey! I'm not sure I understand the question as you mention intermittent 400 errors, DNS cache, and 403 errors. It's possible these are not related issues at all, so it would be best to pick one and get that resolved.

The 403 will leave something in the server's Apache logs at /etc/apache2/logs/error_log. Can you check that file while refreshing the page the see if you find any useful details?

 

[leonep]

right..
the 403 error was because I was enabling debug logging for a domain, but to do that you need to use nginx-debug service
(https://support.cpanel.net/hc/en-us/articles/1500005129641-How-To-Run-The-cPanel-provided-NGINX-In-Debug-Mode-)

for the original error i.e. the 400 on different sites and from different clients cpanel support advised me to uninstall nginx and provide the apache logs with evidence of error

now error disappear ! i am waiting

 

[leonep]

uninstall nginx 400 errors disappear , i wait long time no errors on logs
now i have reinstalled nginx and errors 400 come back again

62.19.xxx.xxx - - [10/Jan/2023:11:32:27 +0100] "POST /start_ctr/login HTTP/2.0" 400 347 "https://domainl.tld/" "Mozilla/5.0 (Linux; An

 

[cPRex]

I would say it's time to make a ticket on this so we can take a look directly.

 

[leonep]

Hello,
after very long ticket i have some useful info.

first of all it must be said that it was a big problem not having the way to reproduce the problem.

another difficulty was the lack of logs because apparently this 400 error does not come out in the warn log level.
moreover I was not able to put the "debug" log level on nginx, not even through cpanel support which inform me about an internal case
https://support.cpanel.net/hc/en-us/articles/12771140611479

in the end, through an automatic page refresh on a site and waiting several minutes, I was able to reproduce the problem.....

once I discovered this I started doing random tests to understand at least where to start working and I discovered that by disabling keep-alive the error disappears!

now I wonder if I can change keep-alive connection and timeout to make sure that re-enabling keep-alive doesn't recreate the 400 error.