..dead ext3 journals, was: Nagios on Debian

[Arnt Karlsen <arnt@c2i.net>]

On Wed, 3 Sep 2003 08:22:22 -0700, 
"Jeremy T. Bouse" <jbouse@debian.org> wrote in message 
<[🔎] 20030903152222.GA23194@UnderGrid.net>:

> On Wed, Sep 03, 2003 at 04:55:53PM +0200, Arnt Karlsen wrote:
> > On Tue, 2 Sep 2003 09:48:40 -0700, 
> > "Jeremy T. Bouse" <jbouse@debian.org> top posted in message 
> > <[🔎] 20030902164839.GC20588@UnderGrid.net>:
> > 
> > > Nagios itself is necessary... Also I've worked out the
> > > dependencies so that Nagios can be setup in a distributed fashion
> > > as documented in the Nagios docs where you only need the central
> > > server to have a web server installed... It didn't make sense to
> > > have it "recommend" the plugins but "depend" on the web server
> > > which meant you couldn't install nagios as a probe-only server
> > > with no web interface and installing Nagios without the plugins
> > > doesn't make much sense at all...
> > 
> > .."a web server" == "an Apache webserver"?  There are _several_ out 
> > there, even some lightweight...
> >
> 
> 	Well it's put as "apache | apache-ssl | httpd" which should
> allow any web server package to be used however the packaging at this
> time only knows how to try and configure itself to run on apache or
> apache-ssl...

..ah, so with boa or monkey webservers I just have to figure out 
config stuff myself.  ;-)

> 	By changing debian/control so that it "Suggests" the web server
> rather than depends on it you could setup a distributed monitoring
> system with Nagios as listed on the Nagios documentation[1]. Also it
> made more sense to change the "Depends" to "nagios-plugins |
> netsaint-plugins (>= 1.2.9.4-7)" rather than as a "Suggests" or
> "Recommends" as Nagios doesn't make much sense without them... 
> 
> 	Regards,
> 	Jeremy
> 
> [1]http://nagios.sourceforge.net/docs/1_0/distributed.html 

..thanks, while I got your attention:  there _are_ ways to monitor 
ext3 and other journalling fs'es for journalling failures?  Had a 
few ext3 go ro on /var and /home and trigging a kernel panic is 
usually overkill, especially on raid-1, I would have expected the 
"journal demon" bail out the "bad disk" and do fsck and restart 
the journal and resync etc, but with Red Hat 7.3-9 and Debian, it's 
"at best" "errors=remount-ro".

-- 
..med vennlig hilsen = with Kind Regards from Arnt... ;-)
...with a number of polar bear hunters in his ancestry...
  Scenarios always come in sets of three: 
  best case, worst case, and just in case.