|
Newbie here. I've got 3 servers and I'm trying to setup LVS to load
balance apache and mysql. No heartbeat at this time.
Our hosting provider doesn't allow the use of internal networks within
their network so I have to do everything with public IP's. (which from
what I've read, isn't that big a deal.) All three servers have 1 eth
port and 5 IP's and all are on the same network: 70.80.90.*
I'm running RHEL4 (2.6.9-42.0.3.ELsmp) on all machines and I'm just
getting a huge headache reading about all this "ARP Problem" stuff.
I've got ldirectord running on LB1 and ipvsadm shows everything
correctly. The main IP on LB1 (70.80.90.5) is bound to eth1 and on lo:0
on the two slaves. I've got "net.ipv4.ip_forward = 1" on LB and set to 0
on slaves.
Care to show the ipvsadm -L -n output?
I ran the following on the two slaves:
arptables -A IN -j DROP -d 70.80.90.5
arptables -A OUT -j mangle -o eth1 -s 70.80.90.5 \
--mangle-ip-s 70.80.90.10 (.15 on slave2)
The preferred way of dealing with this is by instrumenting
arp_{announce,ignore} in the proc-fs.
As far as I can tell, that's all I need to do right? Well, I'm missing
something then because when I open my browser and goto 70.80.90.5,
nothing happens. It seems that LB1 isn't forwarding the requests to the
slaves.
Can you tcpdump on the director? Are you sure there's not some filtering
of illicit traffic on switch ports on your ISP's side?
Regards,
Roberto Nibali, ratz
--
echo
'[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq' | dc
|
