[Linux-cluster] LVS, not so fun today...

[David Brieck Jr.]

On 11/7/06, isplist at logicore.net <isplist at logicore.net> wrote:
> I posted about this before but, no replies so still looking for help.
>
> ---
>
> Don't know why, maybe I've taken on too many things at once but this is just
> confusing me now. Here's all the info I can think of as I ask for help once
> again from the fine folks in this list :).
>
> Same problem, all weekend, trying various combinations. Seems simple enough
> but nothing has worked for me so far. There's something that's just not
> sinking into my brain here. I've included my lvs.cf at the end of this
> message.
>
> LVS0's real eth0 IP is 192.168.1.52.
> LVS1's real eth0 IP is 192.168.1.53.
> CWEB92 is the only real server with the virtual IP of 192.168.1.150 on it's
> ether0 for testing. It's real IP is 192.168.1.92.
> When I ping 150, I reach either LVS servers, depending on who's master but the
> CWEB machine never responds to 192.168.1.150.
>
> So, again, I have 192.168.1.150 as a virtual IP on the first NIC of each REAL
> web server. I have 192.168.1.150 installed on LVS0 as the VIP for the real
> servers and real servers configured.
>
> Nov  6 14:00:49 lb52 pulse[2498]: STARTING PULSE AS MASTER
> Nov  6 14:00:49 lb52 pulse: pulse startup succeeded
> Nov  6 14:01:07 lb52 pulse[2498]: partner dead: activating lvs
> Nov  6 14:01:07 lb52 lvs[2522]: starting virtual service HTTP active: 80
> Nov  6 14:01:07 lb52 nanny[2527]: starting LVS client monitor for
> 192.168.1.150:80
> Nov  6 14:01:07 lb52 lvs[2522]: create_monitor for HTTP/cweb92 running as pid
> 2527
> Nov  6 14:01:07 lb52 nanny[2528]: starting LVS client monitor for
> 192.168.1.150:80
> Nov  6 14:01:07 lb52 lvs[2522]: create_monitor for HTTP/cweb93 running as pid
> 2528
> Nov  6 14:01:07 lb52 nanny[2529]: starting LVS client monitor for
> 192.168.1.150:80
> Nov  6 14:01:07 lb52 lvs[2522]: create_monitor for HTTP/cweb94 running as pid
> 2529
> Nov  6 14:01:12 lb52 pulse[2524]: gratuitous lvs arps finished
> Nov  6 14:01:13 lb52 nanny[2527]: READ to 192.168.1.92:80 timed out
> Nov  6 14:01:13 lb52 nanny[2528]: READ to 192.168.1.93:80 timed out
> Nov  6 14:01:13 lb52 nanny[2529]: READ to 192.168.1.94:80 timed out
> Nov  6 14:01:25 lb52 nanny[2528]: READ to 192.168.1.93:80 timed out
> Nov  6 14:01:25 lb52 nanny[2527]: READ to 192.168.1.92:80 timed out
> Nov  6 14:01:25 lb52 nanny[2529]: READ to 192.168.1.94:80 timed out
>
> serial_no = 88
> primary = 192.168.1.52 (Real, physical IP of the first LVS server)
> service = lvs
> backup_active = 1
> backup = 0.0.0.0
> heartbeat = 1
> heartbeat_port = 539
> keepalive = 6
> deadtime = 18
> network = direct
> nat_nmask = 255.255.255.255
> debug_level = NONE
> monitor_links = 0
> virtual HTTP {
>      active = 1
>      address = 192.168.1.150 eth0:1
> (IP that LVS responds to, then sends connections to read servers, right?)
>      vip_nmask = 255.255.255.0
>      port = 80
>      send = "GET / HTTP/1.0rnrn"
>      expect = "HTTP"
>      use_regex = 0
>      load_monitor = none
>      scheduler = wlc
>      protocol = tcp
>      timeout = 6
>      reentry = 15
>      quiesce_server = 0
>      server cweb92 {
>          address = 192.168.1.92
>          active = 1
>          weight = 1
>      }
>      server cweb93 {
>          address = 192.168.1.93
>          active = 1
>          weight = 1
>      }
>      server cweb94 {
>          address = 192.168.1.94
>          active = 1
>          weight = 1
>      }
> }
>
>
>
> --
> Linux-cluster mailing list
> Linux-cluster at redhat.com
> https://www.redhat.com/mailman/listinfo/linux-cluster
>

Can you post the results of "arping 192.168.1.150"? You might need to
specify the interface and source IP, see the man page for more info.