rc and smf

Bill Hacker wbh at conducive.org
Thu Feb 24 13:39:05 PST 2005


219.GW7819 at xxxxxxxxxxxxxxxx>
In-Reply-To: <20050224211219.GW7819 at xxxxxxxxxxxxxxxx>
Content-Type: text/plain; charset=us-ascii; format=flowed
Content-Transfer-Encoding: 7bit
Lines: 57
NNTP-Posting-Host: 218.253.82.16
X-Trace: 1109281145 crater_reader.dragonflybsd.org 716 218.253.82.16
Xref: crater_reader.dragonflybsd.org dragonfly.kernel:7828

Dan Melomedman wrote:

> Matthew Dillon wrote:
> 
>>    I think what Bill is trying to say, not very diplomatically, is
>>    that the truely important pieces of software out there in world
>>    don't rely on simple-stupid little monitoring programs to deal 
>>    with failures.  They do far more sophisticated tests and the consequences
> 
> 
> Nobody argues that. All I was trying to do is point out benefits to
> simple, stupid little monitoring programs that restart services when
> they fail. Most people seem to have trouble seeing these benefits. You
> have the chance to add this little, in my opinion very useful option to
> Dragonfly's init (not necessarily by adapting runit), or maybe simply
> document it in the Handbook. Why not, really?
> 

Dan, you 'mind me of my late Grandpa.

When winning an argument, he changed sides.

When losing an argument, he changed sides.

His only goal was that the excercise not end 'til the Bushmills was all 
gone.

> 
>>    What I am saying here is that when one is building a highly reliable
>>    system, there's a lot more to it then writing a little service restarter.
> 
> 
> I agree, but as a sysadmin, using a stupid, simple supervisor is the least you
> can do to improve its services uptime. It's very disappointing to see so many
> people disagree with me.

Make you a deal, if yer up to it:

I'll go off and see if I can simplify, readability-fy and 
maintainability-fy a revamp of hoary old /etc/rc, and friends that 
doesn't break anything.

You go off and see about doing the same with the DJB-ified (or any 
other) init toolset of your choice or invention.

Meet you back here in two weeks with *tested and documented results*.

Meanwhile we BOTH cut down our traffic on THIS LIST by  - say 75%?

Happy to exchange views one-on-one in the interim, but let these guys do 
*their* work.

Fair?

Bill







More information about the Kernel mailing list