|
Hi,
You seem to have read part of the Howto :)
========
About ARP info( by run the command arp, is it right ?)
========
Yes, or 'ip neigh show'
=======
About ipvs info
=======
[root@VS1 root]# ipvsadm
IP Virtual Server version 0.8.1 (size=65536)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 210.200.2.5:http wlc
-> 192.168.0.2:http Route 1 0 0
-> 192.168.0.3:http Route 1 0 0
Why did you choose LVS-DR instead of LVS-NAT?
I think I have disabled the arp on the real servers for VIP address,is it
right ?
I don't know if you did, but for LVS-DR this is one way of addressing
the ARP Problem.
I use NAT and the gateway is 192.168.0.1.I enabled the ip_forward on
Cluster. All real servers can connect to internet.But the client can't
connect to anyone of real servers.When I try to connect to the 210.200.2.5,I
found the following information on real server1(192.168.0.2)
[root@RS1 root]# netstat -nt
Active Internet connections (w/o servers)
Proto Recv-Q Send-Q Local Address Foreign Address State
tcp 0 0 210.200.2.5:80 219.218.10.17:38167
SYN_RECV
on the client,
[root@client root]# telnet 210.200.2.5 80
Trying 210.200.2.5...
telnet: connect to address 210.200.2.5: Connection timed out
I can via cluster's 80 port from client if I run the apache on cluster. I
can also via the real server's 80 port only from cluster and LAN.So I think
the firewall's configuration should be ok.
Could you tcpdump the whole path? Oh wait, you're using LVS-DR and have
the DGW of your RS pointing to the director? If so you need to check out
[1].
Would someone give me some hints ?
Thanks !
HTH!
[1] http://www.ssi.bg/~ja/#lvsgw
Best regards,
Roberto Nibali, ratz
--
echo '[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq'|dc
| 