|
Hi,
I see a strange behaviour if I configure the fallback server on the director. I
always get a TCP-RST from one real server if all servers are down:
ldirectord.cf:
virtual=192.168.10.140:25
real=192.168.10.137:25 gate
real=192.168.10.138:25 gate
fallback=localhost:25
ipvs -L
TCP 192.168.10.140:smtp wlc persistent 600
-> localhost:smtp Local 1 0 0
-> 192.168.10.137:smtp Route 0 0 2
-> 192.168.10.138:smtp Route 0 0 0
tcpdump of the communication:
1) Client request:
00:27:13:64:63:ba > 08:00:27:92:13:72, ethertype IPv4 (0x0800), length 74:
192.168.10.143.33833 > 192.168.10.140.25: Flags [S], seq 2111034864, win 5840,
options [mss 1460,sackOK,TS val 14078063 ecr 0,nop,wscale 6], length 0
2) The director forwards the request
08:00:27:92:13:72 > 08:00:27:51:7e:48, ethertype IPv4 (0x0800), length 74:
192.168.10.143.60121 > 192.168.10.140.25: S 244790748:244790748(0) win 5840
<mss 1460,sackOK,timestamp 14116651 0,nop,wscale 6>
3) The real server denies it:
08:00:27:51:7e:48 > 00:27:13:64:63:ba, ethertype IPv4 (0x0800), length 54:
192.168.10.140.25 > 192.168.10.143.33833: Flags [R.], seq 0, ack 2111034865,
win 0, length 0
Please note that ...:ba is the MAC of my client, ...72 is the MAC of my
director and ...48 is the MAC of my first real server.
sysctl of the real server:
sysctl -p
net.ipv4.conf.eth0.arp_ignore = 1
net.ipv4.conf.eth0.arp_announce = 2
net.ipv4.conf.all.arp_ignore = 1
net.ipv4.conf.all.arp_announce = 2
Why does the REAL server answer this connection? Why does the director forward
this connection to the real server.
Thanks for any hints.
--
Dr. Michael Schwartzkopff
Guardinistr. 63
81375 München
Tel: (0163) 172 50 98
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/
LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
or go to http://lists.graemef.net/mailman/listinfo/lvs-users
|
