Reading about that Nextcry-Php issue I decided to follow the instructions given here. I edited my configuration correspondingly.
A restart of nginx failed, as the php-handler was not defined.
So my guess was after comparing my old config to the new one, that fastcgi_pass could be the troublemaker and I add this line of my old config instead of the one listed in the instructions:
So my question is, why is a direct “call” to php-handler not working in my setup? I know this is a random question, as you don’t know my config, but maybe someone stumbled into the same issue and might want to give me a hint.
Oh and more important is to know if this setting is “safe” or could I still be hit by Nextcry with that funky fastcgi_pass parameter?
Thanks for your answer. My ngnix seems to starts up fine with and without the upstream php-handler option that you posted -I did not look into this deeply, but I see that some nc-users are using a configuration that actually use unix:var… instead of 127.0.0.1. Getting more and more confused
You only need an upstream block, if you don’t directly specify a server. So you can just keep your config.
Unix: means that a unix socket is used, which is like a file that can be at different locations. In debian the recommended path recently changed from /var/run to /run