Hello: a small but important update on this issue: I finally got it working: The problem was that server1 was on a different host (XEN server) then the loadbalancer and server2. Perhaps there is a problem with bridging and XEN server (version 4). Moving the server1 vps to the same VPS Host as the load balancer and server 2 fixed my problem and LVS is working perfectly!
I don't know how to fix this on the XEN host servers but perhaps someone else has this issue and know how to fix this. A small but important note for LVS users on VPS systems. Kind regards, Michiel -------- Original Message -------- Subject: Re: [lvs-users] question about LVS From: [email protected] To: [email protected] Date: 05/04/2010 09:35 PM >> On Tue, 2010-05-04 at 20:30 +0200, [email protected] wrote: >>>> Please post the bash script you're using to setup the loopback device >>>> etc. so i can take a look and maybe adjust it. >>> >>> echo "0">/proc/sys/net/ipv4/ip_forward >>> cat /proc/sys/net/ipv4/ip_forward >>> echo 1> /proc/sys/net/ipv4/conf/eth0/arp_ignore >>> echo 2> /proc/sys/net/ipv4/conf/eth0/arp_announce >>> echo 1> /proc/sys/net/ipv4/conf/all/arp_ignore >>> echo 2> /proc/sys/net/ipv4/conf/all/arp_announce >>> /sbin/ifconfig lo:110 194.145.200.87 broadcast 194.145.200.87 netmask >>> 0xffffffff up >>> /sbin/ifconfig lo:110 >>> /sbin/route add -host 194.145.200.87 dev lo:110 >>> >>> I also run this script on server2 in exact the same way. >> >> Looks okay to me. > > Ok..thanks for checking. > >> >>> When I stop Iptables orflush them the reaction from the client is >>> different(as stated before). >>> >>> With iptables on: >>> [r...@knox01 ~]# telnet 194.145.200.87 25 >>> Trying 194.145.200.87... >>> telnet: connect to address 194.145.200.87: No route to host >>> telnet: Unable to connect to remote host: No route to host >>> [r...@knox01 ~]# telnet 194.145.200.87 25 >>> >>> With iptables off (service iptables stop) on server1: >>> (takes a long time) >>> Nothing happens: >>> r...@knox01 ~]# telnet 194.145.200.87 25 >>> Trying 194.145.200.87... >> >> Just realising something. Have you configured the mailserver service on >> the failing realserver to listen on _all_ interfaces? To be sure, >> compare the output of: >> >> $ netstat -tulpen | grep 25 > > Both servers are the same (0.0.0.0:25) > >> >> on both realservers. >> >> >>> Could I run some tcpdumps on real servers and LVS dr to check what is >>> going wrong? >> >> Running tcpdump never hurt anyone :) >> > What kind of tcpdump should I do on the real servers and perhaps load > balancer? > (a tcpdump -n | grep *dest ip* | grep smtp gives too much output) > Here is where my expertise lacks.. :( > >> >> Léon >> > > Michiel > >> > > > > _______________________________________________ > Please read the documentation before posting - it's available at: > http://www.linuxvirtualserver.org/ > > LinuxVirtualServer.org mailing list - [email protected] > Send requests to [email protected] > or go to http://lists.graemef.net/mailman/listinfo/lvs-users _______________________________________________ Please read the documentation before posting - it's available at: http://www.linuxvirtualserver.org/ LinuxVirtualServer.org mailing list - [email protected] Send requests to [email protected] or go to http://lists.graemef.net/mailman/listinfo/lvs-users
