CentOS - Sep 2015 - Keepalived vrrp problem

Hey guys,

 I'm trying to install keepalived 1.2.19 on a centos 6.5 machine. I did an
install from source.

And when I start keepalived this is what I'm seeing in the logs. It's
reporting that the VRRP_Instance(VI_1) Now in FAULT state.

Here's more of that log entry:

Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:  VRRP Instance = VI_1
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Using VRRPv2
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Want State MASTER
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Runing on device
= eth1
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Gratuitous ARP
repeat = 5
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Gratuitous ARP
refresh repeat = 1
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Virtual Router ID
= 51
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Priority = 101
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Advert interval 1 sec
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Accept disabled
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:    Virtual IP = 1
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]:      10.40.116.34/32
dev eth1 scope global
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]: Using LinkWatch
kernel netlink reflector...
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_vrrp[44943]: VRRP sockpool:
[ifindex(3), proto(112), unicast(0), fd(10,11)]
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]: ------<
Global definitions >------
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]:  Router ID
= USECLSNDMNRDBA
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]:  VRRP IPv4
mcast group = 224.0.0.18
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]:  VRRP IPv6
mcast group = ff02::12
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]: ------<
SSL definitions >------
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]:  Using
autogen SSL context
Sep 29 12:06:58 USECLSNDMNRDBA Keepalived_healthcheckers[44942]: Using
LinkWatch kernel netlink reflector...
*Sep 29 12:06:59 USECLSNDMNRDBA Keepalived_vrrp[44943]: Kernel is
reporting: interface eth1 DOWN*
*Sep 29 12:06:59 USECLSNDMNRDBA Keepalived_vrrp[44943]: VRRP_Instance(VI_1)
Now in FAULT state*


And if I do an ifconfig command I see no evidence of an eth1 existing. Also
I can't ping the virtual address that I'm trying to create:

# ping -c 5 10.40.116.34
PING 10.40.116.34 (10.40.116.34) 56(84) bytes of data.
>From 10.40.116.30 icmp_seq=2 Destination Host Unreachable
>From 10.40.116.30 icmp_seq=3 Destination Host Unreachable
>From 10.40.116.30 icmp_seq=4 Destination Host Unreachable
>From 10.40.116.30 icmp_seq=5 Destination Host Unreachable
--- 10.40.116.34 ping statistics ---
5 packets transmitted, 0 received, +4 errors, 100% packet loss, time 14001ms
pipe 3

Here are my configs starting with the first machine:

# cat keepalived.conf
vrrp_instance VI_1 {
interface eth1
state MASTER
virtual_router_id 51
priority 101                   # 101 on master, 100 on backup

vrrp_unicast_bind 10.40.116.30   # Internal IP of this machine
vrrp_unicast_peer 10.40.116.31   # Internal IP of peer
virtual_ipaddress {
 10.40.116.34
}

And here's the config on the second machine:

# cat /etc/keepalived/keepalived.conf
vrrp_instance VI_1 {
interface eth1
state MASTER
virtual_router_id 51
priority 100                   # 101 on master, 100 on backup
vrrp_unicast_bind 10.40.116.31   # Internal IP of this machine
vrrp_unicast_peer 10.40.116.30   # Internal IP of peer
virtual_ipaddress {
 10.40.116.34
}


Does anyone have any experience in solving this kind of problem? Any
suggestions on how to resolve this would be great.

Thanks,
Tim
-- 
GPG me!!

gpg --keyserver pool.sks-keyservers.net --recv-keys F186197B

Tim Dunphy wrote:
> Hey guys,
>
>  I'm trying to install keepalived 1.2.19 on a centos 6.5 machine. I did
an
> install from source.
>
> And when I start keepalived this is what I'm seeing in the logs.
It's
> reporting that the VRRP_Instance(VI_1) Now in FAULT state.
<snip>
Two stupid questions:
  1. Why are you running 6.5, rather than 6.7?
  2. On the host, does it *have* an eth0, or is it, perhaps, em1?

        mark "it's all emx on Dell R720's"

On 09/29/2015 09:14 AM, Tim Dunphy wrote:
> And if I do an ifconfig command I see no evidence of an eth1 existing.
"ifconfig -a" will show you all of your interfaces.

Em 29-09-2015 15:03, Gordon Messmer escreveu:
> On 09/29/2015 09:14 AM, Tim Dunphy wrote:
>> And if I do an ifconfig command I see no evidence of an eth1 existing.
>
> "ifconfig -a" will show you all of your interfaces.
Maybe there is a confusion here. Sounds like Tim thought keepalived 
would create that eth1, like a tunnel interface, but it won't. You have 
to specify an interface that actually exists so that the VIP address 
will be added as a secondary address to ip to that interface.

HTH

   Marcelo