I had the following configuration file ( lvs.cf ) primary = 192.168.2.4 service = lvs rsh_command = rsh backup_active = 0 backup = 0.0.0.0 heartbeat = 1 heartbeat_port = 539 keepalive = 6 deadtime = 18 network = nat nat_router = 192.168.1.5 eth0:0 virtual UnoVirtual { active = 1 address = 192.168.2.5 eth1:0 port = 8900 persistent = 300 send = "GET / HTTP/1.0\r\n\r\n" expect = "HTTP" load_monitor = rup scheduler = rr protocol = tcp persistent = 300 timeout = 6 reentry = 15 server real8900 { address = 192.168.1.45 active = 1 weight = 1 } } virtual DosVirtual { active = 1 address = 192.168.2.5 eth1:1 port = 80 persistent = 300 send = "GET / HTTP/1.0\r\n\r\n" expect = "HTTP" load_monitor = rup scheduler = rr protocol = tcp persistent = 300 timeout = 6 reentry = 15 server Real80 { address = 192.168.1.50 active = 1 weight = 1 } server Linux1 { address = 192.168.1.1 active = 1 weight = 1 } } when i run the pulse as daemon only add a first virtual server bu if I run pulse -n all virtual server are added, why.
It looks like you are mapping the same ip address to two different interfaces (eth1:0 and eth1:1). You cannot do that. It's like trying to define two ethernet cards with the same IP address - there's no routing logic that can be applied. You etiher need to make them two different addresses, or specify the same interface.
Followup.... I not this does not explain the difference in behavior. I'll have to try that and see why.
Any progress?
Another note... We are about to release a new piranha (today hopefully) that has patches which may affect this problem. Please keep a watch on ftp://people.redhat.com/kbarrett/HA/experimental/ and let me know if this makes any difference.
Is this still a problem with the current release on the ftp site?
Status changed until customer verifies that this is still a problem with current release.