I have been using qdPM v9.0 (from qdpm/core/apps/qdPM/config/app.yml) on Centos 6.7 server with PHP 7.0.6 and apache 2.2.x and MariaDB 5.5.x for over a year now without any issues. It seems to be using legacy Symfony 1.4.
I tried to install Let's Encrypt SSL certificates and this upgraded Apache/httpd to 2.2.15, no change in PHP or MariaDB versions.
Upon restarting httpd after SSL certificate installation, suddenly I get 500 Internal Server Error and the httpd error log shows:
...
[Wed Aug 09 14:55:22 2017] [error] [client x.x.x.x] Empty response header name, aborting request
[Wed Aug 09 14:55:32 2017] [error] [client x.x.x.x] Empty response header name, aborting request
...
Also, this is not a misconfiguration of SSL/Apache because other apps on other sub-domains continue work fine, both with and without Let's Encrypt SSL certificates.
Google does not help except some German discussion suggests to use PHP 5.3: https://www.php.de/forum/webentwicklung/php-frameworks/1508593-installation-symfony-framework
Symfony 1 geht nur mit maximal PHP 5.3... Deswegen sagte ich doch hol dir Symfony 3!!!
I cleared the cache several times. I removed all Let's Encrypt SSL configuration as well as restored old self-signed SSL certificates and restored Apache configuration to earlier working state.
And because we take backups daily, I even restored the entire code backup from a few hours before.
This should definitely have worked.
I still get the same error and no clue / hint as to how to debug it. Symfony logging documentation is for its current version, not for 1.4
What could have caused this issue?
How do I enable debugging so I could find where the error "Empty response header name" is being created so that maybe I can patch it?