Re: keepalived with ipv6 - HowTo ?!

 Hi,

 keepalived-announce is not the best list to ask this.

 But, keepalived currently only supports IPv6 for the VRRP part. The 
 IPVS part is not (yet) implemented. So it's not useful to do 
 loadbalancing with it unless you configure it to use external scripts to 
 set up the virtual servers.

 Regards,

 Sander

 On Tue, 4 Jan 2011 13:06:48 +0000, "Zeller, Jan (ID)" 
 <[email protected]> wrote:
> Dear list,
>
> i try to setup keepalived v1.2.1 with IPv6 but it does not work :(
>
> keepalived.conf (on master)
>
> vrrp_instance VI_1 {
>  state MASTER
>  interface eth0
>  virtual_router_id 1
>  priority 100
>
>  authentication {
>  auth_type PASS
>  auth_pass password
>  }
>
>  virtual_ipaddress {
>   myprefix:dead:beef:0:1/64 dev eth0 label eth0:0
>  }
> }
>
> virtual_server myprefix:dead:beef:0:1 3128 {
>        lb_algo dh
>        lb_kind DR
>        protocol TCP
>
>        real_server myprefix:5054:ff:fe66:7122 3128 {
>                TCP_CHECK
>                {
>                 connect_port 3128
>                 connect_timeout 5
>                }
>        }
>
>        real_server myprefix:5054:ff:feab:9f2a 3128 {
>                TCP_CHECK
>                {
>                 connect_port 3128
>                 connect_timeout 5
>                }
>        }
> }
>
> in /var/log/messages I can see the following :
>
> Jan  4 13:26:15 Keepalived_vrrp: Registering Kernel netlink reflector
> Jan  4 13:26:15 Keepalived_vrrp: Registering Kernel netlink command 
> channel
> Jan  4 13:26:15 Keepalived_vrrp: Registering gratutious ARP shared 
> channel
> Jan  4 13:26:15 Keepalived_vrrp: Opening file
> '/opt/keepalived-1.2.1/etc/keepalived/keepalived.conf'.
> Jan  4 13:26:15 Keepalived_healthcheckers: Registering Kernel netlink
> reflector
> Jan  4 13:26:15 Keepalived_healthcheckers: Registering Kernel netlink
> command channel
> Jan  4 13:26:15 Keepalived_healthcheckers: Opening file
> '/opt/keepalived-1.2.1/etc/keepalived/keepalived.conf'.
> Jan  4 13:26:15 Keepalived_healthcheckers: Configuration is using :
> 9361 Bytes
> Jan  4 13:26:15 Keepalived_vrrp: Configuration is using : 33818 Bytes
> Jan  4 13:26:15 Keepalived_vrrp: Using LinkWatch kernel netlink 
> reflector...
> Jan  4 13:26:15 Keepalived_healthcheckers: IPVS: Module is wrong 
> version
> Jan  4 13:26:15 Keepalived_healthcheckers: IPVS: Module is wrong 
> version
> Jan  4 13:26:15 Keepalived_healthcheckers: Using LinkWatch kernel
> netlink reflector...
> Jan  4 13:26:15 Keepalived_healthcheckers: Activating healtchecker
> for service [0.0.0.0:0]
> Jan  4 13:26:15 Keepalived_healthcheckers: Activating healtchecker
> for service [0.0.0.0:0]
> Jan  4 13:26:16 Keepalived_vrrp: VRRP_Instance(VI_1) Transition to
> MASTER STATE
> Jan  4 13:26:16 Keepalived_healthcheckers: TCP connection to
> [0.0.0.0:3128] failed !!!
> Jan  4 13:26:16 Keepalived_healthcheckers: Removing service
> [0.0.0.0:0] from VS [0.0.0.0:3128]
> Jan  4 13:26:16 Keepalived_healthcheckers: IPVS: No such destination
> Jan  4 13:26:16 Keepalived_healthcheckers: TCP connection to
> [0.0.0.0:3128] failed !!!
> Jan  4 13:26:16 Keepalived_healthcheckers: Removing service
> [0.0.0.0:0] from VS [0.0.0.0:3128]
> Jan  4 13:26:16 Keepalived_healthcheckers: IPVS: No such destination
> Jan  4 13:26:16 Keepalived_healthcheckers: Lost quorum 1-0=1 > 0 for
> VS [0.0.0.0:3128]
> Jan  4 13:26:17 Keepalived_vrrp: VRRP_Instance(VI_1) Entering MASTER 
> STATE
>
> What am I doing wrong here ?! The realservers are of course IPv6
> enabled and reachable.
> The same setup works fine with IPv4. Any hints or man pages related
> to keepalived and IPv6 out there ?
>
> OS is Ubuntu 10.10
>
> # ipvsadm -v
> ipvsadm v1.25 2008/5/15 (compiled with popt and IPVS v1.2.1)
>
> ---
>
> Jan
>
> 
> ------------------------------------------------------------------------------
> Learn how Oracle Real Application Clusters (RAC) One Node allows 
> customers
> to consolidate database storage, standardize their database
> environment, and,
> should the need arise, upgrade to a full multi-node Oracle RAC 
> database
> without downtime or disruption
> http://p.sf.net/sfu/oracle-sfdevnl
> _______________________________________________
> Keepalived-announce mailing list
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/keepalived-announce


------------------------------------------------------------------------------
Learn how Oracle Real Application Clusters (RAC) One Node allows customers
to consolidate database storage, standardize their database environment, and, 
should the need arise, upgrade to a full multi-node Oracle RAC database 
without downtime or disruption
http://p.sf.net/sfu/oracle-sfdevnl