It seems that the netstat "bus error" message is generally associated with
heavy net load, which we were experiencing at the time. Several people
recommended using etherfind, truss (or trace) on netstat to examine the net
load or where the command itself is bombing. Also, Don mentioned that the
Sunsolve database made a reference to a kernal patch to possibly fix this
problem. I havent had a chance to try any of the above solutions as the
network has not been overloaded lately. Although a 8 port card in our hub did
go bad over the weekend....Hmmm...
Anyway ..many thanks to:
don@alaska.opensys.com (Don Lenamond)
poffen@San-Jose.ate.slb.com (Russ Poffenberger)
Claus Assmann <ca@informatik.uni-kiel.de>
Richard Sands <ras@hubris.cv.com>
Original Posting:
---------------------------------------------------------------------------
Hi guys.
Sorry if you got this message previously in unfinished form. I clicked the
wrong button before I was finished writing it :-P
As I was saying previously...
Question:
Has anyone ever seen netstat (with no parameters) return the
message "bus error"?
Background:
We are having some network problems this morning
and I have been trying to track down the problem. We are experiencing
heavy packet loss, high collision rates, and as much as 15 secs of a delay
in even displaying what you type back to the screen (if logged in on a
machine over the network). The problem would come and go sporadically.
I did the usual and checked the few remaining thinnet lines, reset the hubs,
DEMPRs, and MPTRs. This would alleviate the problem for about 5 mins and then
it would happen all over again. I began to check other routes...such as
seeing whether the problem was with the server or not. I was using netstat on
our server (a Sparc 10 with pedastal) and the command would work fine
at times when the response on the server was fine, but when response was
lagging, I got the "bus error" message. Anyone ever seen this happen? If so,
does it mean something on the server is going bad or is it just a s/w error
message?
Additional info:
Network is all Suns (mostly 10s, a few 20s, One SS2, and One IPX)mainly connected via aui but with about 5 still on thinnet. After about 2 hours of
seeing this problem. The problem magically went away. I rebooted the server
just to be sure and even did a few prob-scsi and probe-scsi-all just to make
sure that all cards were still visible, but nothing turned up sour.
Can anyone shed some light on this beastie? My network problems are gone
for now, but I am wondering about whether the "bus error" message is a bad
omen or a warning sign. Basically I just need a
1) yes. its a software message. Dont worry about it.
2) yes. its a hardware message. You should worry about it.
3) yes. its a software message. Do worry about it.
4) yes. its a hardware message. Dont worry about it.
5) the reason you got the message was...blah blah blah"
As always...thanks
Greg Harrison
greg.harrison@analog.com
Analog Devices
7910 Triad Center Drive
Greensboro, NC 27409
----- End Included Message -----
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:23 CDT