Simon Horman wrote: > On Wed, Mar 24, 2010 at 03:49:24PM +0200, Deniss wrote: >> ipvs local node setup is broken since kernel 2.6.28 > > Unfortunately I'm not having any luck reproducing this on 2.6.28.
i am running 2.6.28-r9 kernel any additional info u would like to get ? > >> test setup is as following: >> >> MASTER/RS1 (IP 10.0.0.20/24, VIP 10.0.0.1): >> eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state >> UNKNOWN qlen 2000 >> link/ether 00:e0:4c:b5:c7:a2 brd ff:ff:ff:ff:ff:ff >> inet 10.0.0.20/24 brd 10.0.1.255 scope global eth0 >> inet 10.0.0.1/32 scope global eth0 >> FWM 5 wlc >> -> 10.0.0.20:3306 Local 1 0 0 >> -> 10.0.0.30:3306 Route 1 0 0 >> >> RS1 ~ # sysctl -a | egrep "\.(rp_fi|ip_forw)" >> net.ipv4.conf.all.rp_filter = 0 >> net.ipv4.conf.all.arp_filter = 0 >> net.ipv4.conf.all.arp_announce = 2 >> net.ipv4.conf.all.arp_ignore = 1 >> net.ipv4.conf.all.arp_accept = 0 >> net.ipv4.conf.default.rp_filter = 0 >> net.ipv4.conf.default.arp_filter = 0 >> net.ipv4.conf.default.arp_announce = 2 >> net.ipv4.conf.default.arp_ignore = 1 >> net.ipv4.conf.default.arp_accept = 0 >> net.ipv4.conf.lo.rp_filter = 1 >> net.ipv4.conf.lo.arp_filter = 0 >> net.ipv4.conf.lo.arp_announce = 2 >> net.ipv4.conf.lo.arp_ignore = 1 >> net.ipv4.conf.lo.arp_accept = 0 >> net.ipv4.conf.eth0.rp_filter = 1 >> net.ipv4.conf.eth0.arp_filter = 0 >> net.ipv4.conf.eth0.arp_announce = 2 >> net.ipv4.conf.eth0.arp_ignore = 1 >> net.ipv4.conf.eth0.arp_accept = 0 >> net.ipv4.ip_forward = 0 >> >> RS1 ~ # iptables -L -nvt mangle >> Chain PREROUTING (policy ACCEPT 34M packets, 3959M bytes) >> pkts bytes target prot opt in out source >> destination >> 0 0 MARK tcp -- * * 0.0.0.0/0 >> 10.0.0.1 tcp dpt:3306 MARK xset 0x5/0xffffffff >> >> RS2 (IP 10.0.0.30/24): >> lo: <LOOPBACK,UP,LOWER_UP> mtu 16436 qdisc noqueue state UNKNOWN >> link/loopback 00:00:00:00:00:00 brd 00:00:00:00:00:00 >> inet 10.0.0.1/32 scope global lo:1 >> eth0: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast state >> UNKNOWN qlen 2000 >> link/ether 00:17:31:8d:58:3d brd ff:ff:ff:ff:ff:ff >> inet 10.0.0.30/24 brd 10.0.0.255 scope global eth0 >> >> from MASTER trying to connect to VIP (local process tries to connect >> to VIP on local nic): >> telnet -b 10.0.0.20 10.0.0.1 3306 (no connect) >> >> and tcpdump show (tcpdump -i eth0 -n -e -s0 -v host 10.0.0.1): > > Are you sure these packets are showing up on eth0? > I'm seeing this traffic on lo. probably this traffic is on lo as well, but according to 2.6.28 ipvs changes it passed to ipvs code and balanced to RS2 via eth0 > >> 19:22:05.809708 00:e0:4c:b5:c7:a2 > 00:17:31:8d:58:3d, ethertype IPv4 >> (0x0800), length 74: (tos 0x10, ttl 64, id 57858, offset 0, flags [DF], >> proto TCP (6), length 60) 10.0.0.20.42779 > 10.0.0.1.3306: S, cksum >> 0xc2bd (correct), 4115412076:4115412076(0) win 32792 <mss >> 16396,sackOK,timestamp 1955438 0,nop,wscale 8> >> 19:22:05.809811 00:17:31:8d:58:3d > 00:e0:4c:b5:c7:a2, ethertype IPv4 >> (0x0800), length 74: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto >> TCP (6), length 60) 10.0.0.1.3306 > 10.0.0.20.42779: S, cksum 0xbe1e >> (correct), 920820514:920820514(0) ack 4115412077 win 5792 <mss >> 1460,sackOK,timestamp 1008245313 1955438,nop,wscale 8> >> 19:22:09.647727 00:17:31:8d:58:3d > 00:e0:4c:b5:c7:a2, ethertype IPv4 >> (0x0800), length 74: (tos 0x0, ttl 64, id 0, offset 0, flags [DF], proto >> TCP (6), length 60) 10.0.0.1.3306 > 10.0.0.20.42779: S, cksum 0xba5e >> (correct), 920820514:920820514(0) ack 4115412077 win 5792 <mss >> 1460,sackOK,timestamp 1008246273 1955438,nop,wscale 8> >> >> as can be seen, packets goes through ipvs code and forwarded to RS2 and >> REPLIED by RS2 but then dropped somewhere in iptables/ipvs code ! _______________________________________________ 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
