|
I'm sorry to reintroduce this problem to the list, but it no longer seems to be
specific to Keepalived, as I'm experiencing the same problem with the Heartbeat
send_arp utility.
Recap: With the VIP on the loopback device, I can't ping or otherwise connect
to the primary Director/Server which owns the VIP on eth0. I
see "(incomplete)" in the client's ARP table. If I remove the VIP assigned to
lo, I can instantly start pinging and connecting to the server.
As a workaround I can assign the VIP to lo only while the director is in BACKUP
mode, but this is klunky and one more thing that could go wrong during a
failover.
Using eepro100 driver
Both director/servers are attached to a Linksys hub attached to HP 2324 switch.
RedHat 7.3 with updates
kernel-2.4.20-19.7.um.1.i586.rpm (ultramonkey.org -
includes LVS 1.0.9 and hidden_interface patches)
ipvsadm-1.21-1.rh.7.3.um.1.i386.rpm (ultramonkey.org)
keepalived-1.1.1 (keepalived.org)
/ETC/SYSCTL.CONF
net.ipv4.ip_forward = 1
net.ipv4.conf.default.rp_filter = 1
net.ipv4.conf.all.hidden = 1
net.ipv4.conf.lo.hidden = 1
I have the following IP setup:
VIP: 192.168.0.54
RIP: 192.168.0.96 (server1)
RIP: 192.168.0.97 (server2)
SERVER-DIRECTOR1
eth0 - 192.168.0.96
lo - 192.168.0.54
SERVER-DIRECTOR2
eth0 - 192.168.0.97
lo - 192.168.0.54
[root@itdev1 root]# ip addr show
1: lo: <LOOPBACK,UP> mtu 16436 qdisc noqueue
link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00
inet 127.0.0.1/8 brd 127.255.255.255 scope host lo
inet 192.168.0.54/32 scope global lo
2: eth0: <BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast qlen 100
link/ether 00:01:80:02:20:18 brd ff:ff:ff:ff:ff:ff
inet 192.168.0.96/24 brd 192.168.0.255 scope global eth0
inet 192.168.0.54/24 scope global secondary eth0
(Original Post)
http://marc.theaimsgroup.com/?l=linux-virtual-server&m=106148323213198&w=2
|
