<br><br><div class="gmail_quote">On Sun, Jun 19, 2011 at 7:20 PM, math <span dir="ltr"><<a href="mailto:lists@saibo.ch">lists@saibo.ch</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im"><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">
hmm it already fails to delete routes, when updating them,..<br>
<br>
can u try it again with fib_metric=flat<br>
</blockquote>
<br></div>
Here the same thing with the configuration:<br>
FIBMetric "flat"<br>
<br>
As far as I see, the results are the same as without FIBMetric "flat".<br>
<br>
Interestingly the wrong route gets deleted by:<div class="im"><br>
KERN: Deleting <a href="http://10.30.95.101/32" target="_blank">10.30.95.101/32</a> via 10.30.95.101<br></div></blockquote><div>notice target == via </div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
<div class="im">
<br></div>
but not by:<div class="im"><br>
KERN: Deleting <a href="http://10.30.95.101/32" target="_blank">10.30.95.101/32</a> via 10.21.33.10<br>
Prefix 32 = Netmask: 255.255.255.255<br>
<br></div>
Maybe the Netmask/Prefix that once is here and once not could be the problem?<br></blockquote><div>i dont think so</div><div><br></div><div>i guess its simpe the difference whether the route is to a direct neighbour, or via another node,.. </div>
<div>i asume the netmask is always /32, but its not diplsayed by route on first type of routes,.</div><div><br></div></div><div>Markus<br></div><div><br></div><div>p.s as i have no access to an OSX system (and if i had, i`m not sure if i`d find enough time)</div>
<div>so i guess i can/will not debug (and fix) this ,..</div><div><br></div><div>but if u want to try yourself, i would try to assist you.</div><div><br></div><div>p.p.s maybe in several months the bsd routing code is completely renewed (as it lacks some features), </div>
<div>and chances are high that this might fix your OSX issues aswell</div><div>(but as actual bsd code runs well on bsd, and afair some/most osX versions, it might continue to fail on your version/configuration)</div>