1 |
On Thu, Jul 29, 2021 at 5:38 PM Martin Vaeth <martin@×××××.de> wrote: |
2 |
> |
3 |
> Rich Freeman <rich0@g.o> wrote: |
4 |
> |
5 |
> My understanding is that qmali contains a "daemon" which does not |
6 |
> daemonize itself. To my knowledge, you can start such a thing |
7 |
> only with daemontools and systemd; probably the start-daemon of |
8 |
> openrc is not powerful enough for this. |
9 |
|
10 |
I imagine you could just run it under nohup or even screen. There are |
11 |
a couple of solutions. |
12 |
|
13 |
> > Second, it sounds like daemontools requires openrc to run. |
14 |
> |
15 |
> No. As I mentioned in another post, systems can run even without |
16 |
> any init-system |
17 |
|
18 |
I didn't say that systems require a service manager. (I'm assuming |
19 |
that is what you mean by an "init-system.") |
20 |
|
21 |
I said that daemontools apparently requires openrc, at least in its |
22 |
default config. That wasn't really my original claim - others pointed |
23 |
this out earlier in the thread. |
24 |
|
25 |
> By the same argument you have given above, a dependency of |
26 |
> daemontools on some other init system (even a || dependency) |
27 |
> is questionable. |
28 |
|
29 |
That is reasonable. However, at this point I'd really question |
30 |
whether it then belongs in the service manager virtual. If it needs |
31 |
to be run by some other service manager, then is it really acting as |
32 |
"THE" service manager? If it requires some other service manager, |
33 |
then it clearly doesn't make sense to have portage uninstall EVERY |
34 |
other service manager on the system if it is installed, because that |
35 |
basically guarantees that it doesn't work. |
36 |
|
37 |
I'm not an expert in daemontools and it sounds like it is fairly out |
38 |
of date in any case, as is qmail. However, it seems like the |
39 |
dependency situation is inconsistent at best. I'm not going to say |
40 |
WHICH solution is most appropriate, but it seems like just about any |
41 |
of them would have prevented this from happening... |
42 |
|
43 |
-- |
44 |
Rich |