|
So I've got my two server LVS setup going with UltraMonkey 3 on a Fedora FC
4 box. My setup has 115 IPs running under IPVS direct routing using the
UltraMonkey Streamline High Availability and Load Balanced setup. After a
time consuming and frustrating configuration period to get both servers not
to both actively running arps for all of the Ips at the same time, I've got
a few other questions concerning configuration for which documents I've
found for are either not very specific or don't provide an answer.
The main problem is that while my configuration is running and switching
back and forth between servers when one dies, according to the examples in
the Ultramonkey 3 docs, it shouldn't be.
1. While both servers appear to be running (I can get proper pages from both
servers either by reloading the sites from different clients or pulling the
network cables on the servers) the results of ipvsadm -L -n do not seem to
coincide properly with the examples listed.
I've commented out all of the other IPs except for the internal IPs and
restarted heartbeat to simplify the matter.
Cluster/Virtual Service IP: 192.168.0.3
Nitehawk's IP: 192.168.0.4
Seahawk's IP: 192.168.0.5
On the server which appears to be master (going by who is accepting) it
looks like:
[root@nitehawk ~]# /etc/ha.d/resource.d/LVSSyncDaemonSwap master status
master running
(ipvs_syncmaster pid: 7194)
[root@nitehawk logs]# ldirectord ldirectord.cf status
ldirectord is stopped for /etc/ha.d/ldirectord.cf
[root@nitehawk logs]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.0.3:80 rr
-> 192.168.0.5:80 Route 0 0 0
-> 192.168.0.4:80 Local 0 0 0
-> 127.0.0.1:80 Local 1 0 0
On the server not running the ldirector:
[root@seahawk ha.d]# /etc/ha.d/resource.d/LVSSyncDaemonSwap master status
master running
(ipvs_syncmaster pid: 30630)
[root@seahawk ~]# ldirectord ldirectord.cf status
ldirectord is stopped for /etc/ha.d/ldirectord.cf
[root@seahawk ~]# ipvsadm -L -n
IP Virtual Server version 1.2.1 (size=4096)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP 192.168.0.3:80 rr
-> 127.0.0.1:80 Local 1 0 0
The first problem I see is that both hosts are using the fallback rather
than either actual IP. Other than removing the fallback hosts, I'm not too
sure what is going on.
The other problem I've noticed (which may be a result of the above) is that
occassionally my browser alert a message like: "The connection to the server
was reset while the page was loading". There are no connection reset
messages in /var/log/ha-log, /var/log/ldirectord.log, nor in the Apache
logs.
Configs:
-----------------------------------------------------------------
ha.cf
-----------------------------------------------------------------
# File to write debug messages to
debugfile /var/log/ha-debug
#
logfile /var/log/ha-log
#
keepalive 1
#
deadtime 10
#
warntime 2
#
initdead 60
#
udpport 694
bcast eth1 # Linux
#
ucast eth1 192.168.0.4
#
auto_failback on
# node nodename ... -- must match uname -n
node 192.168.0.4 nitehawk
node 192.168.0.5 seahawk.zu.com
#
ping 192.168.0.4
#
respawn hacluster /usr/lib64/heartbeat/ipfail
# deadping - dead time for ping nodes
deadping 10
# defaults to zero
debug 1
#
apiauth ipfail uid=hacluster gid=haclient
-----------------------------------------------------------------
haresources
-----------------------------------------------------------------
# actual configuration for cluster
seahawk.zu.com \
ldirectord \
LVSSyncDaemonSwap::master \
IPaddr2::192.168.0.3/24/eth1/192.168.0.255
-----------------------------------------------------------------
ldirectord.cf
-----------------------------------------------------------------
# Global Directives
checktimeout=3
checkinterval=1
#fallback=127.0.0.1:80
autoreload=no
logfile="/var/log/ldirectord.log"
#logfile="local0"
quiescent=no
# internal virtual host for all sites
virtual=192.168.0.3:80
real=192.168.0.5:80 gate
real=192.168.0.4:80 gate
fallback=127.0.0.1:80 gate
service=http
request="index.html"
receive="Test Message"
scheduler=rr
#persistent=600
#netmask=255.255.255.255
protocol=tcp
checktype=negotiate
__________________________________________________________
Dan Brown
zu.com communications inc. http://www.zu.com
Web Design - Databases - Creative
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
324 Duchess Street Saskatoon SK S7K 0R1
tel.1.306.653.4747 fax.1.306.653.4774
|
