|
Hi,
I am having trouble adding a new real server to our LVS setup using ipvsadm and
am hoping someone here might be able to help.
We have a bunch of existing real servers set up that work fine, but when we try
and add a new one they refuse to work. However, if we reuse one of the existing
ones it works fine.
Below is the output of ipvsadm (anonymised for security hopefully).
The specific thing I am trying to achieve is to add the web.anon.com:1240
realserver and have it direct traffic to web-10.webcluster.anon.com:www.
As it stands it does not work, but if I move web-10 onto web.anon.com:1235 and
remove web-7 from there it works fine and I can access the website running on
web-10.
IP Virtual Server version 1.0.1 (size=65536)
Prot LocalAddress:Port Scheduler Flags
-> RemoteAddress:Port Forward Weight ActiveConn InActConn
TCP web.anon.com:mysql wrr
-> db-1.webcluster.anon.com:mysql Masq 1024 15 168
TCP web.anon.com:3316 wrr
-> localhost:3316 Local 1 0 0
TCP di.anon.com:1231 wrr
-> di-1.webcluster.anon.com:www Masq 10 0 0
TCP web.anon.com:1231 wrr
-> web-1.webcluster.anon.com:www Masq 1024 0 0
TCP web.anon.com:rmtcfg wrr
-> localhost:rmtcfg Local 1 0 0
TCP web.anon.com:1232 wrr
-> web-2.webcluster.anon.com:www Masq 1024 1 2
TCP web.anon.com:1233 wrr
-> web-3.webcluster.anon.com:www Masq 1024 0 0
TCP web.anon.com:1234 wrr
-> web-4.webcluster.anon.com:www Masq 1024 0 0
TCP web.anon.com:1235 wrr
-> web-7.webcluster.anon.com:www Masq 1024 0 0
TCP di.anon.com:1234 wrr
-> di-4.webcluster.anon.com:www Masq 1024 0 0
TCP di.anon.com:1232 wrr
-> di-2.webcluster.anon.com:www Masq 1024 0 0
TCP web.anon.com:1240 wrr
-> web-10.webcluster.anon.com:www Masq 1024 0 0
TCP cy.anon.com:www wlc
-> cycle-1.webcluster.anon.com:www Masq 512 0 0
TCP web.anon.com:www wlc
-> web-7.webcluster.anon.com:www Masq 128 19 182
-> web-1.webcluster.anon.com:www Masq 1024 129 1345
-> web-4.webcluster.anon.com:www Masq 1024 141 1317
TCP nnd.anon.com:www wlc
-> webdirect-1.webcluster.anon.com:www Masq 512 0 2
-> webdirect-4.webcluster.anon.com:www Masq 512 0 1
TCP di.anon.com:www wlc
-> di-2.webcluster.anon.com:www Masq 1024 0 5
-> di-1.webcluster.anon.com:www Masq 10 0 1
-> di-4.webcluster.anon.com:www Masq 10 0 1
TCP svc.anon.com:www wlc
-> svc-1.webcluster.anon.com:www Masq 512 0 0
-> svc-3.webcluster.anon.com:www Masq 512 0 1
Here are the commands I used to add the real server initially:
$ ipvsadm -A -t web.anon.com:1240 -s wrr
$ ipvsadm -a -m -w 1024 -t web.anon.com:1240 -r web-10.webcluster.anon.com:www
Because our existing real servers work fine, I wonder if we have reached some
kernel limit? Here are our kernel settings:
CONFIG_IP_VS_TAB_BITS=16
CONFIG_IP_VS_RR=y
CONFIG_IP_VS_WRR=y
CONFIG_IP_VS_LC=y
CONFIG_IP_VS_WLC=y
CONFIG_IP_VS_LBLC=y
CONFIG_IP_VS_LBLCR=y
CONFIG_IP_VS_DH=y
CONFIG_IP_VS_SH=y
CONFIG_IP_VS_FTP=y
Alternatively, can anybody suggest any further avenues of investigation?
Software version numbers:
Linux kernel: 2.4.18 (Debian)
IPVS: 1.0.1
ipvsadm: 1.21release2-1
ldirectord 0.4.9.0l-7.2
heartbeat 0.4.9.0l-7.2 (we use this to switch ldirectord between redundant
servers)
Any help you can give will be much appreciated. Any further information you
need let me know.
Thanks in advance,
--
o=| Mark |=o
"For Great Justice."
|
