Re: depend take 2

[Rory Toma]

> 
> Sorry, I should have elaborate more instead of throwing it out like
> that; Well the first thing I noticed was the 1 cycle wait to start a
> dependant process. I know you mentioned something about this before
> but at the time I wasn't thinking much about it, that is, until now
> when I have tried it now for myself :) I have short of manages to fix
> the problem (I think?) with this code which is a bit faster in
> control.c:atomic_control_process:
> 
>   if(is(action, "start")) {
>     ASSERT(p->start);
>     if(! is_process_running(p)) {
>       int max_tries= 10;
>       log("%s: (%s) %s\n", action, p->name, p->start->arg[0]);
>       spawn(p, p->start);
>       p->visited= TRUE;
>       while(max_tries--) {
>       if(is_process_running(p))
>           break;
>       sleep(1);
>       }
>     }

This will probably work fine, except maybe instead of a sleep(1), we
want to make it sleep(Run.polltime/max_tries), that way if you have
slow-starting processes, you can tune it with setting the poll time.



> Another problem I found was that if I run 'monit start' from the
> console and both processes is up and running, monit for some reason
> kills one. This is what I have found so far and I just want to see if
> it's easily fixable :-) Agree?

I did not know about this one. I'll test it out.

In the meantime, I'll go ahead and add the above code in.
-- 
Rory Toma               address@hidden
VP of Run Level 5       http://www.trs80.net
Digeo Digital           http://www.digeo.com

signature.asc
Description: This is a digitally signed message part