|
Hi -
First, please email me privately at cristinew@xxxxxxxxxxxxxxxx - our mail
server is in flux and responses to the list may or may not make it through
to me.
I have successfully setup a LVS DR test bed with Mandrake (kernel 2.2) just
recently. I have one director, 2 web servers as real servers and one
client. They are all on a stand-alone hubbed network. In trying to add a
Red Hat 7.1 (kernel 2.4-2) real server (www) to the mix, I'm having loads of
trouble.
Config -
Director
VIP x.x.x.4
DIP x.x.x.11
1st web server - x.x.x.10
2nd web server - x.x.x.15
client - x.x.x.205
Adding the Red Hat box (web server)
x.x.x.16
When the RH box is rebooted, ipchains will reply to ipchains -L with no
problem. After the rc.lvs_dr script is run (unsuccessfully at this point,
it can't ping the VIP during the script run, but will from the command
line), ipchains -L gives me "Incompatible with this kernel".
Questions I can think of --
1) How do I tell whether or not the Red Hat people have already applied the
arp patch?
2) How do I tell whether the redirect arp method has already been applied?
3) How do I go about identifying the actual problem? I didn't have any of
this behavior with 2.2 at all.
4) Why is ipchains' behavior changing so drastically just because of
rc.lvs_dr?
Thanks in advance-
Cristine
--
Cristine Collins
cristinew@xxxxxxxxxxxxxxxx
Duct tape is like the Force. It has a light side & a dark side, and it holds
the universe together.
|
