[ULTRAMONKEY-USERS] Issues with 2-Node HA/LB

Tue Apr 1 05:24:44 EST 2008

I am attempting to set up my first cluster.  My goal is to turn a few
old iPaq's into a simple little web server.  I'm very new to this, but
followed closely along with the UltraMonkey configuration at
http://www.ultramonkey.org/3/topologies/sl-ha-lb-eg.html.  I have also
tried the suggestions at
http://www.jedi.com/obiwan/technology/ultramonkey-rhel4.html which have
gotten me much closer than before.

However, things are still not working right.  If I boot one of the
servers while the other is off, the server will assume the role of
master director and will pick up the virtual IP that I've assigned.  As
soon as I turn the other server on, it also thinks that it is the master
director and attempts to work under the virtual IP as well.  My HTTPD
logs on each server show that the other server is getting the
ldirectord.html file that I've specified in ldirectord.cf, so the
servers should know that the other is alive.

The other strange thing is that even when I manually tell one server to
not be the master (using /etc/init.d/LVSSyncDaemonSwap master stop), the
other server remains master, but it never directs any web requests to
the other server.  I set the server weights in ldirectord.cf to be
50/50, but it seems that whichever is the master always takes the full
load.

I am running Fedora 8 on both.  The reason for this is actually because
the PCs don't have CD or DVD drives and I couldn't manage to get my
hands on one.  I did have another PC that I was playing with Fedora 8
on, so I set it up as a PXE server and used it to install Fedora on the
other guys.  Both of the write-ups I followed used RHEL 4, so I don't
know if that could be part of the problem.  I also had trouble
installing the UltraMonkey RPMs (I honestly can't remember what problems
I had - it was a very long week ago), so I ended up installing heartbeat
with yum.  I think that this could be another potential problem, since I
might be using different versions of heartbeat and ldirectord than the
people who wrote those tutorials.

My configuration is with two PCs with the following IPs, though if I get
this working I would like to add several other identical realservers to
the cluster.

SFTFluster02: 10.50.1.212

SFTFluster03: 10.50.1.213

The virtual IP that I'm using to test is 10.50.1.207.

Here are my config files (with comments removed).

# /etc/ha.d/ha.cf

logfacility local0

keepalive 1

deadtime 10

warntime 5

initdead 120

udpport     694

mcast eth0 225.0.0.1 694 1 0

auto_failback off

node  SFTFluster02.sub.domain.net

node  SFTFluster03.sub.domain.net

ping 10.50.1.1

respawn hacluster /usr/lib/heartbeat/ipfail

apiauth ipfail gid=haclient uid=hacluster

crm off

# /etc/ha.d/ldirectord.cf

checktimeout=15

checkinterval=5

autoreload=no

logfile="/var/log/ldirectord.log"

quiescent=no

virtual=10.50.1.207:80

        fallback=127.0.0.1:80

        real=10.50.1.212:80 gate 1

        real=10.50.1.213:80 gate 1

        service=http

        request="ldirectord.html"

        receive="Server Up"

        scheduler=rr

        persistent=600

        protocol=tcp

      checktype=negotiate

# /etc/ha.d/haresources

SFTFluster02.sub.domain.net \

      ldirectord::ldirectord.cf \

      LVSSyncDaemonSwap::master \

      IPaddr2::10.50.1.207/24/eth0/10.50.1.255

# /etc/ha.d/authkeys

auth 2

2 sha1 password

# /var/www/html/ldirectord.html

Server Up

# /etc/sysconfig/network-scripts/ifcfg-lo

DEVICE=lo

IPADDR=127.0.0.1

NETMASK=255.0.0.0

NETWORK=127.0.0.0

BROADCAST=127.255.255.255

ONBOOT=yes

NAME=loopback

# /etc/sysconfig/network-scripts/ifcfg-lo:0

DEVICE=lo:0

IPADDR=10.50.1.207

NETMASK=255.255.255.255

NETWORK=10.50.1.0

BROADCAST=10.50.1.255

ONBOOT=yes

NAME=loopback

# /etc/sysconfig/network-scripts/ifcfg-eth0

# Intel Corporation 82557/8/9 Ethernet Pro 100

DEVICE=eth0

BOOTPROTO=static

BROADCAST=10.50.1.255

HWADDR=00:03:47:62:4b:06

IPADDR=10.50.1.213

NETMASK=255.255.255.0

NETWORK=10.50.1.0

ONBOOT=yes

GATEWAY=10.50.1.1

I also ran across something that I wasn't sure of in my /etc/networks
file.  There is a line that reads "link-local 169.254.0.0".  I've tried
the setup with the line removed and with it there, and it seems to make
no difference, but I don't know what it is.

Anyhow, I would really appreciate any advice anyone can give me.  The
nature of my hardware setup sort of forces me to use the Streamlined
HA/LB since each PC only has one NIC (unless there's something I don't
know about...imagine that!).

Thanks!

Branton

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.vergenet.net/pipermail/ultramonkey-users/attachments/20080331/da46944f/attachment.htm 