[prev in list] [next in list] [prev in thread] [next in thread] 

List:       tkined
Subject:    Re: [tkined] invalid Netmask?
From:       Erik Schoenfelder <schoenfr () gaertner ! de>
Date:       1999-09-11 15:57:33
[Download RAW message or body]

Hi,

> hmm, sounds you have devices in your network building an byte-swapped
> and therefore wrong icmp mask-reply.
> are these devices with the 0.255.255.255 netmask reply from the same
> manufacturer ?

Thomas>	due to the big number of devices difficult to answer.

Thomas>	We have a complete Fiber Optic network and currently uses
Thomas>	FOIRL NIC only from Allied Telesyn (except for servers with
Thomas>	fast ethernet). Most used NIC modell is the AT-2450FT, but we
Thomas>	also have some AT-1500 and AT-1700 and a few 2450F cards.  All
Thomas>	with the latest drivers.  Back in the office I will check
Thomas>	which NIC modell the failed machines has.

this could be helpful.

i remember a byte swap problem we had with some synoptics devices,
which had wrong bytesex in unreachable and time-exceeded replies.


however, another thought would be to check with another tool if tkined
is broken (which i don't believe ;-) or the devices are to blame.

tcpdump output would be helpful; eg:

>  schoenfr@spog:schoenfr 6 > tcpdump -n icmp
>  tcpdump: listening on eth0
>  16:46:40.310422 194.45.135.2 > 194.45.135.102: icmp: address mask request
>  16:46:40.310422 194.45.135.102 > 194.45.135.2: icmp: address mask is 0xffffffc0 (DF)

to have a look at the traffic itself.


anyway, i'll add an work-around to nmicmpd.c to byteswap such an
impossible answer (leading zero bits, followed by ones), so this can
be handled as expected.

							Erik
--
!! This message is brought to you via the `tkined & scotty' mailing list.
!! Please do not reply to this message to unsubscribe. To subscribe or
!! unsubscribe, send a mail message to <tkined-request@ibr.cs.tu-bs.de>.
!! See http://wwwsnmp.cs.utwente.nl/~schoenw/scotty/ for more information.

[prev in list] [next in list] [prev in thread] [next in thread]