|
Simon Horman wrote:
> Seventh Sven wrote:
>> Simon Horman wrote:
>>> svensven wrote:
>>>> 28 [61.019] IPVS: lookup/in TCP 10.0.0.3:54590->10.0.0.10:9999 hit
>>>> 29 [61.019] IPVS: Enter: ip_vs_dr_xmit, net/netfilter/ipvs/ip_vs_xmit.c
>>
>> LVS B (backup, own IP address 10.0.0.6):
>> FWM 10 rr
>> -> 10.0.0.6:9999 Local 1 0 0
>
> The thing that is confusing me, is that the connection entry should
> have the real server set to 10.0.0.6, and that is a local address on
> LVS B, and thus ip_vs_null_xmit() should be called.
>
> When I say should, I mean, I think that the code ought to behave
> that way. But clearly it isn't for some reason.
I'm getting a bit confused about the mix of direct routing and the
connection table. The purpose of the connection table should be to
determine which incoming packets to consider for IPVS treatment. With
DR, the connection entry dest addr should be the vip. That seems to be
the case here.
On the other hand, if the purpose of the connection table is to
determine the mapping between a client and a realserver, then it would
make sense to either keep the realserver as the dest addr. Then you'd
need an additional mapping between the client-to-vip and
client-to-real entries. Or put mac addresses in the table..
I'll try to dig a bit more into how the connection table looks when
the connection is established to get a better idea of how this works.
S.
_______________________________________________
Please read the documentation before posting - it's available at:
http://www.linuxvirtualserver.org/
LinuxVirtualServer.org mailing list - lvs-users@xxxxxxxxxxxxxxxxxxxxxx
Send requests to lvs-users-request@xxxxxxxxxxxxxxxxxxxxxx
or go to http://lists.graemef.net/mailman/listinfo/lvs-users
|
