[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Killed wrong proc!
From: |
Ted Zlatanov |
Subject: |
Re: Killed wrong proc! |
Date: |
29 Sep 2004 10:00:57 -0400 |
User-agent: |
Gnus/5.110003 (No Gnus v0.3) Emacs/21.3.50 (gnu/linux) |
On Wed, 29 Sep 2004, architect@webalive.biz wrote:
> Hmm. I think Redhat wants you to use "service httpd start" or
> whatever. There are also a variety of other actions, such as reload and
> status which could be useful. However, unfortunately, the whole thing is
> pretty non-standardised :).
If /etc/init.d doesn't work properly in Redhat, that's news to me.
I stopped using it (for Debian) when the FC switch happened, and was
not aware of a difference.
Maybe cfengine can have an option to do "service" commands, but it
seems silly if /etc/init.d works just fine. What's next, Redhat
will discover daemontools? :)
The reason why I only used "start" and "stop" in my example is
because those are the only standard options to /etc/init.d scripts
AFAIK.
If I'm ignorant of any standards in this regard, I would appreciate
some pointers. Unix startup/shutdown scripts seem to be stable
relative to a SysV hierarchy these days, with almost no pure
BSD-style startup/shutdowns, but I'm not aware of a standard.
Anyhow, I think cfengine should support /etc/init.d scripts as an
option in processes: because that's at least 90% of the Unix systems
out there, and it's better to start a process that way (stop; start)
than sending a signal directly to the process. The signal handling
differs among implementations, for one.
Ted
- Re: Killed wrong proc!, (continued)
- Re: Killed wrong proc!, Ted Zlatanov, 2004/09/22
- Re: Killed wrong proc!, Ted Zlatanov, 2004/09/28
- Re: Killed wrong proc!, Chip Seraphine, 2004/09/28
- Re: Killed wrong proc!, Tim Nelson, 2004/09/28
- Re: Killed wrong proc!,
Ted Zlatanov <=
- Re: Killed wrong proc!, Tim Nelson, 2004/09/30
- Re: Killed wrong proc!, Ted Zlatanov, 2004/09/30
- Re: Killed wrong proc!, Mark Burgess, 2004/09/30