Hi Mikael, The question is: When do you consider your server overloaded? Is that when it is running more than 50 requests at a time? Or is that when it manages 50 sessions from an application point of view?
In the first case, maxconn on the server line statement may help. In the second case, it may be a bit more complicated to do. You can find an example of how you can do this using stick tables here: http://blog.exceliance.fr/2012/09/19/application-delivery-controller-and-ecommerce-websites/ This is the overusage protection section. Baptiste On Wed, Apr 17, 2013 at 2:09 PM, Mikael Kermorgant <[email protected]> wrote: > Hello, > > We are going to launch a web app for a critical subscription step in our > school. A high load the day we open the server, and would therefore like to > protect it from that and setup a waitqueue with ha proxy. > > But we wonder how to setup ha proxy for that. > > Our idea is that : > > * we have 3 servers : ha proxy, the web app's server and a third delivering > a static html page > * if there are too many candidates (let's say over 50), only some of them > will get access to the web app. The other will be redirected to the static > html page. > * if a candidate starts the web process, he must be able to continue until > the end of the procedure (idea : use a cookie to have persistence) > > We have read a bit about maxconn, cookie persistence and maxqueue but we're > not sure if our idea can be built with them. > > Any opinion about that ? > > Best regards, > -- > Mikael Kermorgant
