Hello, Sorry, i replied to the wrong post.
Yes, the http://www.howtoforge.com/loadbalanced_mysql_cluster_debian is the correct link of the tutorial. I have configured the lo:0 interfaces with the loadbalancer ip as described in page 7. on node sql1 i have ip a s lo:0 1: 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 127.0.0.1/8 scope host lo inet 192.168.201.100/32 scope global lo:0 on node sql2 ip a s lo:0 1: 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 127.0.0.1/8 scope host lo inet 192.168.201.100/32 scope global lo:0 I have added another host to the VPN network with the IP 192.168.201.14. >From it, issuing arping gives off the following: arping 192.168.201.100 ARPING 192.168.201.100 42 bytes from fe:ff:ff:ff:ff:ff (192.168.201.100): index=0 time=1.205 msec 42 bytes from fe:ff:ff:ff:ff:ff (192.168.201.100): index=1 time=1.467 msec 42 bytes from fe:ff:ff:ff:ff:ff (192.168.201.100): index=2 time=1.460 msec 42 bytes from fe:ff:ff:ff:ff:ff (192.168.201.100): index=3 time=1.481 msec I also ran tcpdump on the load balancer host and i have noticed a strange thing. When i do tcpdump on the tun0 interface, i get the chatter between the management node and the sql nodes but no traffic directed to the VIP. tcpdump -i tun0 port 3306 tcpdump: WARNING: arptype 65534 not supported by libpcap - falling back to cooked socket tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on tun0, link-type LINUX_SLL (Linux cooked), capture size 96 bytes 09:21:26.929772 IP ip-192-168-201-1.ec2.internal.57553 > ip-192-168-201-6.ec2.internal.mysql: S 3517636247:3517636247(0) win 5840 <mss 1460,sackOK,timestamp 13004102 0,nop,wscale 7> 09:21:26.935957 IP ip-192-168-201-6.ec2.internal.mysql > ip-192-168-201-1.ec2.internal.57553: S 1616136901:1616136901(0) ack 3517636248 win 5792 <mss 1366,sackOK,timestamp 171829194 13004102,nop,wscale 7> & nbsp; 09:21:26.935975 IP ip-192-168-201-1.ec2.internal.57553 > ip-192-168-201-6.ec2.internal.mysql: . ack 1 win 46 <nop,nop,timestamp 13004103 171829194> 09:21:26.936123 IP ip-192-168-201-1.ec2.internal.57553 > ip-192-168-201-6.ec2.internal.mysql: F 1:1(0) ack 1 win 46 <nop,nop,timestamp 13004103 171829194> 09:21:26.936480 IP ip-192-168-201-1.ec2.internal.34946 > ip-192-168-201-10.ec2.internal.mysql: S 3505041773:3505041773(0) win 5840 <mss 1460,sackOK,timestamp 13004103 0,nop,wscale 7> 09:21:26.936710 IP ip-192-168-201-6.ec2.internal.mysql > ip-192-168-201-1.ec2.internal.57553: P 1:128(127) ack 1 win 46 <nop,nop,timestamp 171829199 13004103> 09:21:26.936723 IP ip-192-168-201-1.ec2.internal.57553 > ip-192-168-201-6.ec2.internal.mysql: R 3517636248:3517636248(0) win 0 09:21:26.937079 IP ip-192-168-201-6.ec2.internal.mysql > ip-192-168-201-1.ec2.internal.57553: F 128:128(0) ack 1 win 46 <nop,nop,timestamp 171829199 13004103> On the other hand, if i run tcpdump on the lo interface of the loadbalancer and from another session on the loadbalancer run a "telnet 192.168.201.100 3306" i get: tcpdump -i lo port 3306 tcpdump: verbose output suppressed, use -v or -vv for full protocol decode listening on lo, link-type EN10MB (Ethernet), capture size 96 bytes 09:22:08.367299 IP ip-192-168-201-100.ec2.internal.43644 > ip-192-168-201-100.ec2.internal.mysql: S 3556049393:3556049393(0) win 32792 <mss 16396,sackOK,timestamp 13014460 0,nop,wscale 7> 09:22:08.369408 IP ip-192-168-201-100.ec2.internal.mysql > ip-192-168-201-100.ec2.internal.43644: R 0:0(0) ack 3556049394 win 0 So the traffic appears to be directed at lo instead of tun0. Thanks. --- On Thu, 4/15/10, L.S. Keijser <[email protected]> wrote: From: L.S. Keijser <[email protected]> Subject: Re: [lvs-users] Mysql cluster ldirector IP Can't connect to MySQL To: "LinuxVirtualServer.org users mailing list." <[email protected]> Date: Thursday, April 15, 2010, 9:43 PM On Thu, 2010-04-15 at 12:17 -0700, Bacterie wrote: > ERROR 2003 (HY000): Can't connect to MySQL server on '192.168.201.100' (111) Okay, first of all, the link you pasted was incomplete. I can only assume you meant http://www.howtoforge.com/loadbalanced_mysql_cluster_debian Assuming you configured the loopback adapters on the realservers as lo:0 as described on page 7, could you do an 'ip a s lo:0' and paste the output? Could you also do an 'arping VIP' from a client (note: _not_ the director or any of the realservers) and paste the first couple of lines? -- Léon _______________________________________________ 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
