Ascend Archive
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: (ASCEND) PANIC: Problems with 6.0.0

On Tue, Mar 03, 1998 at 01:51:12PM -0400, Tim Basher wrote:
> > > It appears that the Max suddenly stops routing packets
> > > over established profiles that worked correct a second before. It
> > > seems to be a highlevel issue: using wandsess I can see that the Max
> > > stops sending out any packet on the session, completely ignoring that
> > > I'm pinging to the other side like crazy. Hang up and redial of the
> > > session and all works again, but we've seen this too often now to
> > > call it spurious - it is a systematic problem. The routing table
> > > doesn't show up any problems (Ok I didn't dive into whether the wanX
> > > displayed there is indeed the one associated with the profile...).
> 
> Okay.  How is the profile being created?  Is it from a local profile,
> the answer profile, or a RADIUS profile?  Are you using assigned IP

I've seen it with both local (NVRAM) and RADIUS based profiles. It does
not seem to make a difference.

> addresses, local address pools, or global address pools?  You say

Static adresses. In the cases where I saw it until now, it is straight-
forward routing to remote P50s with official IPs and Networks all on the
way between /24 and /30.

> you do not get response from the ping, what does traceroute show?

Traceroute reaches the Max as the first hop and then goes to the stars.
Is consistent with pinging directly from the Max TermServ, which does
not change anything.

> You say you cannot see the ping, but if the peer pings you do you see
> the packet in the wadsess output?  Were you seeing the wadsess output
> before the problem occured?

I still see RECVs when the problem occurs. The latter, however, is a
really good question. As of now, I've only looked into this when the
problem was already there, and saw RECVs but no XMITs. I did not see
it before, so it is possible that the call actually never XMITed any-
thing and I just assumed it stopped doing so. I assumed that this
happens to calls which are up for a long time because if I detect
such call it is almost always very high in the SNMP sessionActiveTable,
and this means "old" calls. But you are correct, it may be that the
call did never XMIT anything. It's just my feeling that it is the
other way around, based also on:

 - have read similar descriptions already here on the list
 - had a lockup in a download with ISDN @Home that appeared after
   half an hour or so. I did not investigate because I did not
   classify it a failure of the Max at this moment.
 - Had a case where pinging a remote peer succeded and some minutes
   later hung while I was pretty sure the peer did not hang up
   and dial again between the two tests

Anyway, dial traffic on the Max is too high to be able to debug
something in depth.

> Check the output of the terminal server "show users" and "show ip routes"
> commands.  Are there any duplicate addresses, overlapping network routes,
> duplicate names?

show ip routes looked perfect. I did not see anything weird, and the
routes belonging to the profile in question looked the same way as
with operative ones. I haven't looked into "show users" but I have
my session viewer (SNMP) running constantly. It shows the profile in
normal shape, with IP and netmask correct. No duplicates either.

> Check the output of the "show ip stats" and "show icmp" commands.  Are
> any of the error counters increasing when you try to ping the peer?

Damn, that one is still to be tried. Let's see when it happens again
and I'm still an 6.0.0 I'll look this up. "show ip stats" after my
last reset just says:

  1338120 packets received.
       10 packets received with header errors.
        0 packets received with addresss errors.
        0 packets forwarded.

*aehm* you see, ipForwDatagrams is not alone...

        0 packets received with unkown protocols.
        0 inbound packets discarded.
   105354 packets delivered to upper layers.
   110911 transmit requests.
        0 discarded transmit packets.

Maybe it's not logged, maybe it just did not happen yet since the last
reset. I'll see.

-- 

Kanther-Line: PGP SSH IDEA MD5 GOST RIPE-MD160 3DES RSA FEAL32 RC4

+-o-+--------------------------------------------------------+-o-+
| o |               \\\- Brain Inside -///                   | o |
| o |                   ^^^^^^^^^^^^^^                       | o |
| o | Andre' Beck (ABPSoft) beck@ibh-dd.de XLink PoP Dresden | o |
+-o-+--------------------------------------------------------+-o-+
++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>
References: