Re: debugging SLP problems on Mac OS 9/X
Re: debugging SLP problems on Mac OS 9/X
- Subject: Re: debugging SLP problems on Mac OS 9/X
- From: Smith Kennedy <email@hidden>
- Date: Wed, 9 Feb 2005 08:03:30 -0700
Hi Glen,
If they were unique messages being sent out, what fields in the various
headers (IP, UDP, etc.) would be changing each time? If they were the
same packets just being re-cast over the network, which fields would be
the same? I have been looking at some of my networking books, but
haven't found anything yet that tips me off one way or the other.
TIA,
Smith
On Feb 8, 2005, at 7:43 PM, Glenn Anderson wrote:
At 7:15 PM -0700 2/8/05, Smith Kennedy wrote:
I was wondering if anyone might have any insight into a problem we
are having with Mac systems on our site. The problem seems to
involve SLP on both Mac OS 9 and Mac OS X.
I posted a related message to the client-management ML a while ago,
but got no meaningful replies (thanks to John Welch for trying,
though).
A couple of months ago, for several days, there were these "multicast
packet squalls" that would occur periodically, and would be visible
on all our subnets on the site. Each "squall" consists of 250K or so
SLP service request queries being made within 1-2 minutes. All
queries in a particular "squall" would emanate from the same system,
and the system was always a Mac OS 9 or Mac OS X system.
<snip>
My questions are these:
- Does anybody have any hunches why this would be happening?
I have seen something like this happen before, it was some sort of
ethernet switch problem causing a multicast loop. If you are seeing
the same problem, it is not the Mac going crazy, it is one packet from
the Mac being looped back and forth between a couple of ethernet
switches. One way to tell is the packet storm will continue even if
you unplug the Mac that started it. The network where I saw this was
too complex to try and track down the problem in the time available,
we ended up solving it by rate limiting multicast packets on the main
switch. As it was just one packet going 'round and 'round, as soon as
it hit the rate limit the packet was discarded, and this stopped the
storms in a fraction of a second. My guess is it was some sort of
layer 3 IGMP and/or spanning tree problem with one or more ethernet
switches.
Glenn.
Attachment:
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Macnetworkprog mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden