[mrtg-developers] Re: Patch with new option to not skip dead hosts
oetiker at ee.ethz.ch
Thu Apr 10 15:29:10 MEST 2003
Today Andreas Plesner Jacobsen wrote:
> MRTG 2.9.28 has a behaviour which currently makes it unsuitable for our
> purposes; it skips hosts when it has problems getting data for one of
> the interfaces. This happens in two cases for us:
> 1. Seems to be an IOS/MRTG interop issue that makes mrtg unable to get
> the interface data when ifref=descr on a Cisco 10000. I'll investigate
> this further and either submit a bugreport to Cisco or a patch for mrtg.
> 2. When an interface has been torn down. This happens automatically and
> we don't want to do an entire cfgmaker run of all 6000 interfaces each
> This patch introduces a new option for mrtg.cfg: DontSkipDead, which can
> be enabled to let mrtg do further processing of a router even though a
> single interface fails.
The actual idea of this funciton was, that if a whole router was
'offline' it would get skipped. This happens as soon as a single
interface fails. The problem seems to be that the code does not
distinguish between 'timeout' and 'failure message for a single
interface'. So the patch could be even without adding an extra
option, just having it detect propperly that the host is down.
Because what you want to avoid at all cost is that mrtg tries to
poll each and every interface of a deat router ...
______ __ _
/_ __/_ / / (_) Oetiker @ ISG.EE, ETZ J97, ETH, CH-8092 Zurich
/ // _ \/ _ \/ / System Manager, Time Lord, Coder, Designer, Coach
/_/ \.__/_.__/_/ http://people.ee.ethz.ch/~oetiker +41(0)1-632-5286
Unsubscribe mailto:mrtg-developers-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:mrtg-developers-request at list.ee.ethz.ch?subject=help
More information about the mrtg-developers