i wrote a very simple patch for ipvs that enables a kernel config
option that allows to choose where IPVS intercepts incoming
connections. These are the options:
- LOCAL_IN (default: works as usual)
- PRE_ROUTING (puts LVS input right after the mangle PREROUTING and
before the nat PREROUTING chain)
neat. I thought it was hard enough to move that it wouldn't be just
an option :-)
Hehe, yes, it was pretty easy ;)
what we'd really like is ipvs hooked into the FORWARD chain. Can you
do this too?
To be honest i don't understand the reason for hooking LVS into the
FORWARD chain, because this way it would not get the LOCAL_IN traffic
and at the same time it would have the same NAT problem as with the
LOCAL_IN hook. Maybe i'm missing something, but it seems that
PREROUTING is the best point for LVS to act like a real router,
because it gets packets that haven't been NATed yet.
The only negative thing is that traffic can't be filtered in a regular
way, but using fwmark and the mangle table the user can select the
traffic that has to be handled by LVS.
If you like i can also add this option, but i'm not sure if it's
really useful :)
I tried it on some test boxes and it seems to work pretty well,
i'll do some stress testing in the next few days. I could send you
a setup example if you like...
Ok, i'll try to write a short document/example ASAP.
:: e n d i a n
:: open source - open minds
:: raphael vallazza
:: phone +39 0471 631763 :: fax +39 0471 631764
:: http://www.endian.com :: raphael (AT) endian.com
To unsubscribe from this list: send the line "unsubscribe lvs-devel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html