LVS
lvs-users
Google
 
Web LinuxVirtualServer.org

Fwd: DNS Loadbalancing problem

To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Subject: Fwd: DNS Loadbalancing problem
From: James Wilson <james.wilson77@xxxxxxxxx>
Date: Sun, 28 Aug 2005 14:53:36 +0930
Has anyone had a chance to look at this? Or even point me to some 
documention on DNS failover with LVS?

---------- Forwarded message ----------
From: James Wilson <james.wilson77@xxxxxxxxx>
Date: Aug 25, 2005 8:02 PM
Subject: DNS Loadbalancing problem
To: lvs-users@xxxxxxxxxxxxxxxxxxxxxx

Hi All,

We are running Red Hat Enterprise 3 AS with:

ipvsadm-1.21-9.ipvs108
piranha-0.7.6-1

And

Kernel 2.4.21-27.0.2.ELsmp

We have a basic requirement to utilise 2 DNS server behind a pair of
LVS boxes. We noticed that our 8 Squid servers (also behind a pair of
LVS servers) display huge delays retrieving web pages when a DNS
server from /etc/resolv.conf fails. To aliviate this, we decided to
stick the DNS servers behind the LVS boxes.

The problem I find is that I can resolve from the DNS server directly,
however through the LVS box they fail. The following also reports in
/var/log/messages:

Aug 25 19:55:41 xxx nanny[27167]: READ to
172.18.0.121:53<http://172.18.0.121:53>timed out
Aug 25 19:55:41 xxx nanny[27168]: READ to
172.18.0.122:53<http://172.18.0.122:53>timed out


lvs.conf extract (TCP is there not for zone transfers but requests
that are > 512bytes):

virtual DNS-UDP {
active = 1
address = 172.18.0.120 <http://172.18.0.120> sw0:3
vip_nmask = 255.255.255.0 <http://255.255.255.0>
fwmark = 53
port = 53
load_monitor = none
scheduler = wlc
service = none
protocol = udp
timeout = 6
reentry = 15
quiesce_server = 0
server DNS3 {
address = 172.18.0.121 <http://172.18.0.121>
active = 1
weight = 5
}
server DNS4 {
address = 172.18.0.122 <http://172.18.0.122>
active = 1
weight = 5
}
}

virtual DNS-TCP {
active = 1
address = 172.18.0.120 <http://172.18.0.120> sw0:3
vip_nmask = 255.255.255.0 <http://255.255.255.0>
fwmark = 153
port = 53
load_monitor = none
scheduler = wlc
protocol = tcp
timeout = 6
reentry = 15
quiesce_server = 0
server DNS3 {
address = 172.18.0.121 <http://172.18.0.121>
active = 1
weight = 5
}
server DNS4 {
address = 172.18.0.122 <http://172.18.0.122>
active = 1
weight = 5
}
}


IPTables mangle rules:

-A PREROUTING -d 172.18.0.120 <http://172.18.0.120> -p udp -m udp --dport 53 
-j MARK --set-mark 0x35
-A PREROUTING -d 172.18.0.120 <http://172.18.0.120> -p tcp -m tcp --dport 53 
-j MARK --set-mark 0x99


Can anyone help, or point to some useful documentation -

http://www.austintek.com/LVS/LVS-HOWTO/HOWTO/LVS-HOWTO.services.single-port.html#DNS

Stumped me even more :)


Cheers,

James

--
Senior Systems Engineer
MCP+I, MCSE, RHCE, CCA
CSM Technology Adelaide
ph: 08 8418 7804
fax: 08 8418 7820
email: james.wilson@xxxxxxxxxx
internet: http://www.csm.com.au


-- 
Senior Systems Engineer
MCP+I, MCSE, RHCE, CCA
CSM Technology Adelaide
ph: 08 8418 7804
fax: 08 8418 7820
email: james.wilson@xxxxxxxxxx
internet: http://www.csm.com.au

<Prev in Thread] Current Thread [Next in Thread>