Then there is a web application that has 10-12 pages with multiple post / GT DB calls. When site traffic reaches 1000 or more (concurrent users), there is usually an Apache crash / other problem, which is a small number, we have updated the server with good RAM and resources when our The system admin recommends loading loads on bomb and other custom scripts and suggesting to go away from Apache. Some things do not understand for me as Apache is very wrong, we look at CloudFlare for caching to handle a few thousand concurrent users. Here's what he has suggested:
Replacement with Apache + mod_fcgi with Nginx + php-fpm which can handle the server more users, and then test it. Or 2. For testing: To run the scenario from 10-20 servers is required. In fact, what is needed is a more complex blitz. AOLOog Create a server, which takes all those hours, then clone it in the cloud and multiply by the number of servers required for approximately 1 hour of testing.
Once again several DB calls anf HT access. Does this make the NGX better than APK in this case?
Basically, nginx is an event based, so it is capable to manage more requests concurrently However, as the MySQL DB seems to be a choke point, it is quite possible that NGN will not solve all your problems. Probably moving towards any SSQL type in a database, which is better on scaling horizontally (if it is possible) will help.
No comments:
Post a Comment