1 |
On Sunday 22 October 2006 07:52, Roy Marples wrote: |
2 |
> OK, from my persepective we had RC_STRICT_NET_CHECKING because we couldn't |
3 |
> previously handle multiple provides. Here's how the new way translates. |
4 |
|
5 |
i use it because i want to disable most net based requirements ... my machines |
6 |
all use this: |
7 |
|
8 |
> lo - net is up if lo is up |
9 |
> Just have net.lo in the boot runlevel - it should always be there anyway. |
10 |
|
11 |
so you're saying if net.lo is in the boot runlevel, and i put say net.eth0 |
12 |
[wired] and net.eth1 [wireless] into my default runlevel, i dont have to |
13 |
worry about eth0/eth1 screwing up my net services ? |
14 |
|
15 |
for example, if i start stopping/starting net.eth[01], i certainly dont want |
16 |
sshd/tftpd/rsyncd/etc... starting and stopping as well ... nor do i want |
17 |
these services to fail because net.eth[01] failed to start at boot |
18 |
|
19 |
thus RC_STRICT_NET_CHECKING=lo gave me the perfect behavior |
20 |
-mike |