On Mon, Feb 05, 2007 at 05:57:12PM -0500, Joseph Hardeman wrote:
> Greetings,
>
> We are having a problem, we have setup a LVS cluster for our streaming
> servers and are having problems with the LVS servers. I
> startup/failover the servers and running ipvsadm -L shows active and
> inactive connections but after a few minutes they stop showing any
> connections, but LVS is still routing traffic to the real servers. I
> am using the same setup as on the Web site LVS servers but they are
> not working properly. Also, the real strange thing is that I had our
> image servers on the same system as the streaming media servers and
> while the images stopped sending traffic the streaming servers still
> continued to route properly. I have removed the images from the same
> LVS cluster but still the LVS is not showing the output of the routing
> to the real servers.
>
> Server OS: CentOS 2.6.9-42.0.8.ELsmp
> ipvsadm -v
> ipvsadm v1.24 2003/06/07 (compiled with popt and IPVS v1.2.0)
>
> [root@image-stream-lvs1 ha.d]# ipvsadm -L
> IP Virtual Server version 1.2.0 (size=4096)
> Prot LocalAddress:Port Scheduler Flags
> -> RemoteAddress:Port Forward Weight ActiveConn InActConn
> TCP patch.kaneva.com:http lc persistent 3600
> -> Stream1:http Route 1 0 0
> TCP streaming.kaneva.com:http lc persistent 3600
> -> Stream4:http Route 1 0 0
> -> Stream5:http Route 1 0 0
> -> Stream1:http Route 1 0 0
> -> Stream3:http Route 1 0 0
> -> Stream2:http Route 1 0 0
>
>
> I am hoping someone can help me out. Our web servers lvs is running
> beautifully. Which is so strange.
Perhaps you haven't solved the ARP problem, one of the real-servers
has answered ARP for the VIP, and all the traffic is going directly
to that real-server.
--
Horms
H: http://www.vergenet.net/~horms/
W: http://www.valinux.co.jp/en/
|