> > > > > > wtf I have to create this to find the error. > > > > > > > > > > > > - LOG ERRORS!! > > > > > > > > > > Errors were probably output/logged. > > > > > > > > > > > > > no nothing in error_log, nothing in ssl_error_log > > > > all such tries > > > > /httpd -DFOREGROUND -e debug -E /dev/stderr > > > > > > > > nothing on console > > > > > > > > > > - START EVERYTHING THAT IS GOOD AND JUST FAIL THE VHOST!!!!! > > > > > > > > > > I think it's much worse to allow the server to partially start. > > > > > > > > > > > > > How is having all websites down better than having 1 website > down???? > > > > > > Because you get the appropriate feedback. > > > > > > > I will get that any ways. Make sure you get an internet connection > that shuts down if the latency is above >40ms so you get the proper > feedback. See how you handle that. > > Not really analogous. If you've debugged your issue that silently > failed, report it with specifics and without the histrionics. >
Oh what about this analogy: User A's mobile phone is not configured correctly. So lets turn of everyones phone in the US. It is related to crt loading, something went wrong with auto renewal probably. Nothing else happening there on that server. I think is pretty stupid to have 99% of your websites fail and need to be forced to debug it. In this case it was not even important virtual host failed. Maybe this would have even auto recovered with next crt run. Any way it is just stupid to have an apachectl configtest and then fail on restart. Htf should you test if apache would safely restart.
