> I'm working on load balancing some sqlgrey servers using ipvsadm (non of the
> other bells and whistles, just the director itself).  Anyway, at first glimps
> things look like they are running well until I find that I have a lot of
> connections in ESTABLISHED mode running on the sqlgrey real servers.  Looking
> at the director there are corresponding active connections, and finally,
> looking at postfix, there are only a few active connections.  I have worked
> with the ipvsadm group to ensure arp routing and other items are taken care of
> as per their recommendations.
> 
> What I am seeing is that postfix seems to be sending the close request down
> the pipe to the load balancer and the director isn't responding (properly?),
> postfix is then marking the connection as FIN_WAIT, and finally, the
> connection is being dropped by the postfix server.  The problem appears to be
> that the director isn't either passing the request forward or isn't
> acknowledging it at all.
> 
> Now, when I swing the postfix directly to the IP of one of the real nodes, no
> problem.  So, I was thinking maybe persistence.  So I change director to be
> sticky for 15 minutes, but no go.
> 
> Has anyone seen this behavior before on a load balanced policy server?
> 
> Anyone have any ipvsadm scripts for a policy server they would like to share?
> I'm using wlc nat'ed connections currently.


Things work much better.  The lost connections were because of iptables.  I 
have this rule early on for server that has the director.  I guess the ACK FIN 
is an technically an invalid state...

-A INPUT -p tcp -m conntrack --ctstate INVALID -j LOG --log-prefix "FW-I BF: " 
-A INPUT -p tcp -m conntrack --ctstate INVALID -j REJECT --reject-with 
icmp-port-unreachable

Apr 26 04:36:02 wall1 kernel: FW-I BF: IN=br0 OUT= PHYSIN=eth1 
MAC=00:50:56:b1:63:bc:00:0c:29:92:be:b7:08:00 SRC=10.80.66.24 DST=10.80.55.11 
LEN=52 TOS=0x08 PREC=0x00 TTL=64 ID=40835 DF PROTO=TCP SPT=52114 DPT=3917 
WINDOW=363 RES=0x00 ACK FIN URGP=0 

> Gary

Reply via email to