Issue 503 error on specific page with POST method

[Jérémy S]

Hi everyone,

I'm having errors since I suspended then re-activated my Plesk-hosted website...

The error is specific to a page of my application with POST method. It is a Symfony application that has litterally been working thousands of times before, which is even more frustrating.

I tried everything I found on the Web, including changing the PHP handler, but I get similar errors with the three of them (see the logs below). Even reinstalling the whole Ubuntu+Plesk+application failed.

I desperately don't know what to do with this... Does somebody know a solution?

 

[Bitpalast]

There's nothing wrong with the services, it is a script error. "End of script output before headers" can have a variety of reasons, but the bottom line is: The script does not deliver anything useful to web server.

 

[Jérémy S]

There's nothing wrong with the services, it is a script error. "End of script output before headers" can have a variety of reasons, but the bottom line is: The script does not deliver anything useful to web server.

I've been testing my PHP scripts on my local machine in both dev and prod modes without any error... So I guess this is not the issue here.

However, my logs show the page is executed twice on every request... I guess that's where the issue relies.


This does not happen on my local server... Do you know what could cause this behavior?

 

[Bitpalast]

Duplicate script execution might be caused by redirects, e.g. from non-SSL to SSL and non-www to www spelling of the domain. But again: The root cause can for sure be found in the script or its configuration, not in a bug of the server software.

 

[Jérémy S]

Duplicate script execution might be caused by redirects, e.g. from non-SSL to SSL and non-www to www spelling of the domain. But again: The root cause can for sure be found in the script or its configuration, not in a bug of the server software.

Thank you for your time!

Actually there was no duplicate execution but misleading logging configuration (two mechanisms logging stuff to the same file ).

Mystery remains but I found out there could be an issue with DB/ORM... However it seems to occur in a very precise case, which happens to coincide with a critical part of the application. Nevermind...

Considering these new elements, I guess you're right saying that Apache/nginx are working fine.

The problem is not solved but I learned something today, thank you again!

 

[Jérémy S]

Hi everyone,

I'm having errors since I suspended then re-activated my Plesk-hosted website...

The error is specific to a page of my application with POST method. It is a Symfony application that has litterally been working thousands of times before, which is even more frustrating.

I tried everything I found on the Web, including changing the PHP handler, but I get similar errors with the three of them (see the logs below). Even reinstalling the whole Ubuntu+Plesk+application failed.
View attachment 18778
I desperately don't know what to do with this... Does somebody know a solution?

After three days of intense labour, I finally found out that the source of the error is PHP itself!
Version 8.0.5 of the pdo_pgsql extension has a bug causing a "segmentation fault", which simply aborts everything. I was able to reproduce it on my own computer so there's no doubt now.

I hope the Plesk team will release a version supporting PHP 8.0.6 soon! This is the only fix...

One last time, immense thank you @Peter Debik for putting me on the right tracks!