1 |
>> So shouldn't it wait till the network has started and I have an IP |
2 |
>> address |
3 |
>> from ifplugd? |
4 |
> |
5 |
> Well, isn't the problem here that the network isn't being requested to |
6 |
> start (until ntp tries to make a connection, which of course attempts to |
7 |
> start it, but then it's too late)? |
8 |
|
9 |
[snip] |
10 |
|
11 |
> So I would suggest |
12 |
> |
13 |
> rc-update add net.eth0 default |
14 |
|
15 |
Slightly off... |
16 |
|
17 |
ifplugd is used to bring a network connection up/down when the network |
18 |
device is plugged in or unplugged. So the OP doesn't want net.eth0 starting |
19 |
at boot automatically because the cable might not be plugged in at the time. |
20 |
|
21 |
Unfortunately that's as much as I know about ifplugd. I do know that the |
22 |
gentoo network scripts will ensure that the "provide net" flag is defined |
23 |
that ntp-client relies upon to do it's thing. |
24 |
|
25 |
What I don't know is how integrated ifplugd is with the gentoo network |
26 |
scripts, if it too ensures the "provide net" flag is specified. Nor do I |
27 |
know if, upon cable switching from eth0 to eth1 for example, the network |
28 |
services that were previously tied to eth0 would be restarted to now use |
29 |
eth1... |
30 |
|
31 |
I'm willing to gues that in the OP's case the ifplugd is not setting the |
32 |
"provide net" flag correctly and/or it is setting the flag before a cable is |
33 |
actually connected. In any case it's probably down & dirty with the gentoo |
34 |
networking scripts to figure out how to get the timing to work right... |
35 |
|
36 |
|
37 |
Dave |
38 |
|
39 |
-- |
40 |
gentoo-user@g.o mailing list |