UNIX for Dummies Questions & Answers

Hello guys,
Hope you are doing great!!
Right now I'm using keepalived daemon in order to ensure a Load Balancing function between my two servers.
First of all, I tried using the famous "Round-Robin Scheduling" algorithm and it worked well, but I'm kind of obligated to use the "Source Hashing Scheduling" algorithm, unfortunately I didn't find out an example that explains the usage of this method.

Do you guys have any supplementary information concerning this subject?

Thank you in advance.

It appears to amount to if(x_is_down || x_is_overloaded) use_next_equivalent;

That it's a hash table just means the order they're chosen is indeterminate (but unlikely to change).

Hello again,

Thank you for your reply,

I think I got the idea, and unfortunately it doesn't help resolving my problem, in fact, I have 2 servers and one VIP, and let's say for example I want one specific client (with a unique IP address) to always connect to one specific server, and in general the clients' requests will be distributed on these two servers using the "RR" algorithm, do you know how doing this?

Thank you again !

How did you use round-robin before?

I added this code to the configuration file of keepalived

Code:

virtual_server IP 80 {
 	delay_loop 6
 	lb_algo rr
 	lb_kind DR
 	persistence_timeout 50
 	protocol TCP
real_server FirstserverIP Port {
	TCP_CHECK {
  	connect_timeout 3
	}
  }
  real_server SecondserverIP Port {
	TCP_CHECK {
  	connect_timeout 3
	}
  }


 }

And this can't be used now?

yes and I'm using it right now, but it doesn't help resolving my problem,I want my load balancer to be in "rr" mode for almost all the time exept for one specific client (with a unique IP address), I want this client for example to always connect to one specific server(let's say for example server number 2) .