Re: rc and smf
On Thu, Feb 24, 2005 at 03:49:42PM +0100, Joerg Sonnenberger wrote:
>On Wed, Feb 23, 2005 at 08:18:36PM -0500, George Georgalis wrote:
>> I've been meaning to give runit a spin. I've been very happy with
>> socklog (syslogd replacement) from smarden.org; daemontools and other
>> djb software too, despite the build on the target license requirement.
>>
>> My first impression with Solaris Service Management Facility, is the
>> likely proprietary nature of it. While the license didn't exactly jump
>> out and bite me, I suspect the technology as well as the code is not
>> open. Is it?
>
>I consider most of daemontools useless software, because it is build
>on the wrong assumptions. A normal daemon should not need supervision,
>just like a normal daemon can either write its own log files [if the
>volume is high] or just use syslog.
I run apache under supervise. I use the apache internal logging facility
for each virtual domain access.log, because it does exactly what I
need, but I run a multilog instance for each virtual domain error log,
as that works better for me than error.log. All of this is separate
from the apache daemon logs (restarts etc), which are under multilog as
well as the apache run script. The entire installation is very modular,
and has worked very well, over the years (nearly 2).
I still run services under regular init (sysv) process, and I could
put daemons to respawn in my inittab. I guess this translates to
bsd /etc/ttys but there is a lot more to supervise than restart:
alarms, ulimit and env come to mind; but what I really appreciate
is the overlaying robust infrastructure that also gives unrestrained
flexibility.
>I'm not sure I like the SMF, it has to be carefully evaluted wether
>the increase in parallelism really gains noticable speed and wether
>it compensates for the more confusion log on startup. Always keep
>in mind that higher parallelism can slow down e.g. IO access by
>adding seeks.
valid. I didn't make it through the hype to really understand smf,
but I understand runit does support daemeon dependencies (ie ordering)
and delay to start..
Making a runit port for a base BSD would be a lot of work: code and
commiter philosophy -- I'd be interested, but I certainly don't have
time for it. What I do ask is that it remain an option for users,
so no unusual roadblocks, technical or otherwise, remain for those
who want to use it.
Thanks,
// George
--
George Georgalis, systems architect, administrator Linux BSD IXOYE
http://galis.org/george/ cell:646-331-2027 mailto:george@galis.org
討論串 (同標題文章)
完整討論串 (本文為第 25 之 53 篇)：