1 |
> I believe the author is counting on a total failure of eth0 to drop the |
2 |
> primary route. Once eth0 goes down that route is withdrawn from the |
3 |
> routing table leaving only the eth1 route. Linux without turning on IP: |
4 |
> advanced router in the kernel will use only one gateway. I assume it |
5 |
> uses the first one you config and once that disappears the second will |
6 |
> be used. |
7 |
> The problem is that eth0 will hardly every fail completely. The only |
8 |
> way it would is if it is directly connected to say a DSL device and the |
9 |
> device became unplugged. That would completely drop eth0 and then the |
10 |
> failover would happen. If only the connection upstream from the DSL |
11 |
> device went down then the Linux box would happily send data to the up |
12 |
> ether interface of the DSL, router, switch, etc. |
13 |
> It would also work if you're terminating connections directly into the |
14 |
> Linux box... using a ds-1 card or something like that. Again this |
15 |
> assumes that the connections fail totally. A ds-1 usually will, but in |
16 |
> offices you're sometimes handed ether which terminates on a switch. That |
17 |
> switch will be up, but the router or upstream connection has failed. |
18 |
> I can't find much on the /proc/sys/net/ipv4/route/gc_timeout setting, |
19 |
> but what I did find suggests that it only provides timing as to when the |
20 |
> route is dropped. It doesn't add anyway to monitor the connection. |
21 |
> |
22 |
> It's an interesting little hack assuming I'm right about what he's doing |
23 |
> with it, but I wouldn't go around implementing it on most networks. |
24 |
> |
25 |
|
26 |
I have yet to try this (no test machines), but if you are correct this |
27 |
is pretty much useless to ~90% of those trying to achieve real |
28 |
redundancy on dual links. I didn't find any concrete details regarding |
29 |
gc_timeout's real purpose. If anyone here can easily read linux code |
30 |
they may peruse net/ipv4/route.c which contains the code relevant to |
31 |
gc_timeout. |
32 |
|
33 |
-- |
34 |
Eduardo Tongson |
35 |
<pornadmin.net/~tongson> |
36 |
|
37 |
-- |
38 |
gentoo-server@g.o mailing list |