a few concepts

mixo <[email protected]>
Newsgroups gmane.linux.redhat.piranha
Message-ID <[email protected]>
I have been  trying to setup failover services for a couple of days (on 
and off).
Initially I was using 3 machines, a primary node, a backup node and 1
real (web) server, and then changed to just having having 2, a primary 
node and a server.
In the last  situation, where I had 2 machines ,  the primary always 
took over ("nanny" did not
start either). With the 2 nodes + server, the primary started fine : it 
did not assume the IP
of the  real web server immediately, but soon took over.

What I am trying to do is have a setup where if the web server 
(wilma.example.com/
/192.168.1.4) fails, then the primary node 192.168.1.3 takes over.

What am I missing in my setup? Is using private ip's a a bad idea? Do I 
have to have
a back up node?

(All the machines have public IP's but can't be reached directly as they 
are behind a
firewall, I can't even access those from my machine without going  
through the firewall.
That's why I have resorted to using private IP where I could.)


I have included my logs and configuration :

2 Nodes + a web server
++++++++++++++++++++++/etc/sysconfig/ha/lvs.cf+++++++++++++++++++++++++++++++
serial_no = 77
primary = NODE_1_IP
primary_private = 192.168.1.3
service = fos
backup_active = 1
backup = NODE_2_IP
backup_private = 192.168.1.1
heartbeat = 1
heartbeat_port = 539
keepalive = 18
deadtime = 72
network = direct
reservation_conflict_action = preempt
debug_level = NONE
failover wilma.example.com {
     address = 192.168.1.4 eth0:8
     vip_nmask = 255.255.255.0
     active = 1
     port = 80
     timeout = 60
     send = "GET / HTTP/1.0\r\n\r\n"
     expect = "HTTP"
     start_cmd = "/etc/rc.d/init.d/httpd start"
     stop_cmd = "/etc/rc.d/init.d/httpd stop"
}

======================================================================


The real server:
++++++++++++++++++++++/var/log/httpd/access.log 
--wilma.example.com++++++++++++++
192.168.1.3 - - [24/Jul/2002:10:59:03 +0200] "GET / HTTP/1.0" 200 2947 
"-" "-"
192.168.1.3 - - [24/Jul/2002:10:59:29 +0200] "get i" 400 - "-" "-"
192.168.1.3 - - [24/Jul/2002:10:59:58 +0200] "GET / HTTP/1.0" 200 2947 
"-" "-"
192.168.1.3 - - [24/Jul/2002:11:00:58 +0200] "GET / HTTP/1.0" 200 2947 
"-" "-"

======================================================================

The primary node:
+++++++++++++++++/var/log/messages -- dhcp4.example.com - NODE 
1+++++++++++ ++++++
Jul 24 10:59:28 dhcp4 pulse[3589]: STARTING PULSE AS MASTER
Jul 24 10:59:28 dhcp4 pulse[3589]: Starting Failover Service Monitors
Jul 24 10:59:28 dhcp4 fos[3591]: Stopping local services (if any)
Jul 24 10:59:28 dhcp4 fos[3591]: Shutting down local service 192.168.1.4:80
Jul 24 10:59:28 dhcp4 pulse: pulse startup succeeded
Jul 24 10:59:28 dhcp4 httpd: httpd shutdown failed
Jul 24 10:59:28 dhcp4 fos[3591]: Warning; shutdown of local service 
192.168.1.4:80 returned error 1
Jul 24 10:59:28 dhcp4 fos[3591]: Starting monitor for 192.168.1.4:80 
running as pid 3591
Jul 24 10:59:28 dhcp4 nanny[3619]: Failover service monitor for 
NODE_2_PUBLIC_IP:80 started
Jul 24 10:59:28 dhcp4 nanny[3619]: Remote service 192.168.1.4:80 is 
available
Jul 24 11:00:40 dhcp4 pulse[3589]: partner dead: activating failover 
services
Jul 24 11:00:40 dhcp4 fos[3591]: Shutting down due to signal 15
Jul 24 11:00:40 dhcp4 fos[3591]: Shutting down monitor for 
wilma.example.com 192.168.1.4:80 running as pid 3619
Jul 24 11:00:40 dhcp4 nanny[3619]: Terminating due to signal 15
Jul 24 11:00:40 dhcp4 fos[3591]: will now exit to notify pulse...
Jul 24 11:00:40 dhcp4 fos[3627]: Stopping local services (if any)
Jul 24 11:00:40 dhcp4 fos[3627]: Shutting down local service 192.168.1.4:80
Jul 24 11:00:40 dhcp4 httpd: httpd shutdown failed
Jul 24 11:00:40 dhcp4 fos[3627]: Warning; shutdown of local service 
192.168.1.4:80 returned error 1
Jul 24 11:00:40 dhcp4 fos[3627]: Starting local service 192.168.1.4:80 ...
Jul 24 11:00:41 dhcp4 httpd: httpd startup succeeded
Jul 24 11:00:45 dhcp4 pulse[3626]: gratuitous fos arps finished
=======================================================================

++++++++++++++++++++++/etc/rc.d/rc.lvs+++++++++++++++++++++++++++++++++++++++
#!/bin/sh

iptables -t nat -F -v
echo 1 >/proc/sys/net/ipv4/ip_forward
modprobe iptable_nat
iptables -v -t nat -A POSTROUTING -s 192.168.1.0/24 -j MASQUERADE
ipvsadm -A -t NODE_1_IP:80 -s rr
ipvsadm -a -t NODE_1_IP:80 -r 192.168.1.1:80 -m
ipvsadm -a -t NODE_1_IP:80 -r 192.168.1.2:80 -m
ipvsadm -a -t NODE_1_IP:80 -r 192.168.1.3:80 -m
ipvsadm -a -t NODE_1_IP:80 -r 192.168.1.4:80 -m
ipvsadm -a -t NODE_1_IP:80 -r 192.168.1.5:80 -m

=======================================================================
And,
++++++++++++++++++dhcp4.example.com- NODE 1+++++++++++++++++++++++++++++++
[root@dhcp4 root]# /sbin/ipvsadm -L
IP Virtual Server version 0.9.7 (size=65536)
Prot LocalAddress:Port Scheduler Flags
  -> RemoteAddress:Port           Forward Weight ActiveConn InActConn
TCP  dhcp4.example.com:http rr
  -> 192.168.1.5:http             Masq    1      0          0        
  -> 192.168.1.4:http             Masq    1      0          0        
  -> 192.168.1.3:http             Local   1      0          0        
  -> 192.168.1.2:http             Masq    1      0          0        
  -> 192.168.1.1:http             Masq    1      0          0        

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
lmpx.com only provides a reader for public news (NNTP) servers. It is not affiliated with the servers or forums shown here and is not responsible for the content of articles, which is written by their respective authors.