Re: Some stuff....

[Rory Toma]

BTW, I like the socket idea, and I think that should be used all the
time. Here's a problem I ran into this morning:

monit stop named

I could then not use monit to restart named, as I had allow hostnames in
the monitrc, which could not resolve...

If always had socket, this wouldn't be a problem.

On Wed, 2003-01-08 at 14:08, Christian Hopp wrote:
> On Wed, 8 Jan 2003 address@hidden wrote:
> 
> > I have one additional feature I'd like to see. I'll describe the situation
> > I'm trying to solve and we can come up with a solution.
> > Whilst testing monit, I ran "monit stop sshd". My connection went sway
> > before I could run "monit start sshd". Fortunately for me, this machine
> > was local...
> 
> Weird... if I stop my sshd just the master server is stopped but the
> connection remains.  All "connection servers" are given to init.  You have
> to kill just the master sshd with SIGTERM.  Then it does work... I have
> tried it on Linux and I think I did it several times on Solaris, too.
> 
> > It would be nice to have a flag saying to restart something, even if
> > do_validate is false, after a certain period of inactivity. Perhaps even a
> > mode where we do "monit test", so that monit would reset the flags after a
> > given period.
> 
> Mmmh... the problem is how should monit know that you are cutting your own
> rope?  What if the user intended to stop that certain service and suddenly
> it starts again?
> 
> Christian
-- 
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