[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: routed stepping on bgpd's feet
- To: misc_(_at_)_openbsd_(_dot_)_org
- Subject: Re: routed stepping on bgpd's feet
- From: Arvid Grøtting <arvidg_(_at_)_netfonds_(_dot_)_no>
- Date: Tue, 07 Sep 2004 09:44:35 +0200
- Cancel-lock: sha1:hd07I+XC9zAbFly1jSXWtIoMN7U=
- Mail-copies-to: never
- Mail-followup-to: misc_(_at_)_openbsd_(_dot_)_org
- Organization: No such thing.
Henning Brauer <lists-openbsd_(_at_)_bsws_(_dot_)_de> writes:
> * Arvid Grøtting <arvidg_(_at_)_netfonds_(_dot_)_no> [2004-09-06 19:03]:
> > I guess I'm still in "is it really supposed to be like this?"-mode.
>in this case, it might be, but cannot tell without logs.
Sure. Which logs would you like, for when I reproduce this? (There
will be a lot of log entries; too much to send to the list. Who would
> while I completely agree with you on the fact that sticking with tools
> OpenBSD provides in the base install, I am not so sure wether this is a
> good idea when it comes to routed...
Oh, well. Trouble is, the competition consists of, that I can find:
- zebra/quagga, which is a bit on the heavy side code-size-wise, to
say the least, and
- gated, which has license problems.
> I suspect routed stealing them behind our back. I currently do not see
> why we apparently don't notice tho...
>> $ echo `netstat -rnf inet | wc -l` - `bgpctl sh fib | wc -l` | bc
> can you look at those and try to find a pattern?
The pattern I found is: They're negative, while they should be
positive (and match `arp -an | grep -v static | wc -l`).
>> Running another RIP-2 routing daemon on these particular gateways
>> works around this
> ewp? what do you mean? running two instances of routed fixes it???
Ehm, no. Running a different routing daemon, e.g. gated, works; it's
not what I'd rather do in the long run, though.