[Xorp-users] Interfaces all going "down"
Pavlin Radoslavov
pavlin@icir.org
Tue, 28 Mar 2006 09:12:38 -0800
> Dear All,
>
> I have recently deployed a multicast router
> using Xorp (release candidate for 1.2).
> This seems to run fine for a while,
> and then I noticed problems and
> when I then interrogate it I see...
>
> show pim interface
> Interface State Mode V PIMstate Priority DRaddr Neighbors
> discard0 DISABLED Sparse 2 DR 1 127.127.0.1 0
> eth1 DOWN Sparse 2 NotDR 1 0.0.0.0 0
> eth2 DOWN Sparse 2 NotDR 1 0.0.0.0 0
> eth3 DOWN Sparse 2 NotDR 1 0.0.0.0 0
> eth4 DOWN Sparse 2 NotDR 1 0.0.0.0 0
> register_vif DOWN Sparse 2 DR 1 0.0.0.0 0
>
> [The machine actually has five interfaces, but one, eth0 zero,
> is not connected to a switch at present and so I removed it from
> Xorp config. four interfaces 0-> 3 on one four port card
> and the fifth on a second card - Fedora Core 4 OS].
>
> Any suggestions of a strategy I should adopt to
> try to track the fault that is leading to the interfaces
> going "down". Incidently, the interfaces are still
> shown as "up" with ifconfig at OS level and I can
> log in remotely using ssh just fine.
If you haven't restarted XORP, what is the output of "show
interfaces", "show mfea interface" and "show igmp interface" xorpsh
commands?
Also, did you notice how long time (hours, days?) it took to get
into this state and do you have the xlog output when approximately
this happened.
Finally, do you remember whether the cables connecting the router
have been disconnected and connected back in some stage?
If you don't have the xlog output, then I'd recommend to enable the
traceoptions inside mfea4, igmp and pimsm4 and save it to a file which
can be reviewed later if/when this happens again. You can enable the
traceoptions by adding the following to the mfea4, igmp, pimsm4
configuration blocks:
traceoptions {
flag all {
disable: false
}
}
Pavlin