1 |
Bo Andresen: |
2 |
> On Wednesday 22 March 2006 20:26, Sergio Polini wrote: |
3 |
> > I'ld like to know how the sequence of init scripts is set up. |
4 |
> |
5 |
> Did you look in the Gentoo handbook? |
6 |
> http://www.gentoo.org/doc/en/handbook/handbook-x86.xml?part=2&chap= |
7 |
>4 |
8 |
> |
9 |
> If after reading that you still have unanswered questions then ask |
10 |
> again. |
11 |
|
12 |
I had read that page (perhaps I should have said that I had ;-), but: |
13 |
- before() and after() functions are not mentioned (I can guess that |
14 |
they are "normal" functions defined only in net.*); |
15 |
- the arguments to after and before seem service names (f.i., "but |
16 |
want your service to be started before (or after) another service"), |
17 |
but... what kind of service is "net"? Is it net.lo, net.eth?, |
18 |
netmount? There is no "net" service... |
19 |
|
20 |
I'm using a notebook (HP dv5078) and an all-in-one printer (HP PSC |
21 |
1510). The hplip service has not worked until I removed it from the |
22 |
default runlevel and executed it in local.start. |
23 |
I've found that hplip worked if started after iptables, didn't work if |
24 |
started before iptables. |
25 |
I've contacted HP, and their answer was: |
26 |
> HPLIP requires the loopback interface for TCP/IP socket |
27 |
> communication. I would expect the following order at boot time |
28 |
> for proper operation. |
29 |
> |
30 |
> iptables |
31 |
> loopback interface |
32 |
> hplip |
33 |
> cups |
34 |
|
35 |
My default init sequence [was] is different: |
36 |
- domainname |
37 |
- net.lo |
38 |
- syslog-ng |
39 |
- anacron |
40 |
- net.eth0 |
41 |
[- hplip, when it didn't work] |
42 |
- cupsd |
43 |
- gpm |
44 |
- iptables |
45 |
- netmount |
46 |
- local -> hplip |
47 |
|
48 |
I'ld like to try the sequence suggested by HP, but I can't understand |
49 |
what could I do. For example, how could I execute iptables (which is |
50 |
executed "before net", but after net.*) before net.lo? |
51 |
|
52 |
> > Moreover, I'd like to know if it's possible to verify the actual |
53 |
> > sequence of the init scripts. I can look at the commands that a |
54 |
> > make command would execute but do not execute them, or at the |
55 |
> > problems reported by fsck avoiding to repair them, by the -n |
56 |
> > option. Could I check an init sequence in a similar way? |
57 |
> |
58 |
> I don't think so. But it will check syntax of the entire script if |
59 |
> you run any command on it. Only syntax though. |
60 |
|
61 |
Yes, I could try: |
62 |
- replacing "before net" with "before net.lo" in iptables; |
63 |
- adding "after iptables" to cupsd; |
64 |
and so on, but even if some editing worked, eventually after a few |
65 |
booting failures, that would be a "blind workaround" ;-) |
66 |
|
67 |
Cheers |
68 |
Sergio |
69 |
-- |
70 |
gentoo-user@g.o mailing list |