|
Hi !
I want to create a cluster of 2 Apache web servers. What I need is a way to
send 100 % of the load on the first real server and switch the exedent
requests locally.
Ex : Server1 (172.17.20.151) gets 100 % of the load until it reaches 400
connections. The next requests would go locally on 127.0.0.1.
ipvsadm -C
ipvsadm -A -t 172.17.20.161:0 -s rr -p 300
ipvsadm -a -t 172.17.20.161:0 -r 127.0.0.1:0 -w 0
ipvsadm -a -t 172.17.20.161:0 -r 172.16.20.151:0 -w 100
ipvsadm -l
IP Virtual Server version 0.9.5 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 172.17.20.161:(null) rr persistent 300
-> localhost:(null) Local 00 0 0
-> 172.17.20.160:(null) Route 100 0 0
I could rewrite:
ipvsadm -e -t 172.17.20.161:0 -r 127.0.0.1:0 -w 100
ipvsadm -e -t 172.17.20.161:0 -r 172.17.20.151:0 -w 0
when I see 400 ActiveConn using a script...
But I dont want the request on 127.0.0.1 to be persistent !
Any idea on how to do this ?
Can I desactivate the persitent connections on 127.0.0.1 ?
Any better way of doing this ?
Where in the code should I look to create my own sets of scheduling algo ?
Thanks in advance.
Jean-Francois Nadeau
| 