Hi ALl;
Thanks to everyone for who answered. The consensus was a Serial Driver bug was
being triggered. Seems the real solution will be an OS upgrade and PPP upgrade.
I applied the patch 100513-04 (I wasn't able to find one other that seemed
like it might be useful). This actually made the problem worse. I had to back
it out because the PPP link wouldn't stay up for more than 20 seconds.
Until I get time to upgrade, I guess I am out of luck. Thanks for the help.
I have attached my Original Question and a few of the useful replies.
Thanks to:
----------
cliff@iscmed.med.ge.com
Kevin.Sheehan@uniq.com.au
A.J.Cole@cbl.leeds.ac.uk
dhaut@level1.com
pag@gpf.scg.boulder.co.us
-Jim
-----------------------------------------------------------------------------
Jim Murff, NIS (murff@nicimg.com) Voice # (619)635-8678
Nicolet Imaging Systems Inc., San Diego, CA. Corp # (619)695-6661
Senior Software Engineer/System Admin. Fax # (619)695-9902
-----------------------------------------------------------------------------
----- Begin Included Message -----
To: ppp-users@morningstar.com, sun-managers@eecs.nwu.edu
Subject: Help with PPP crashing sun.
Hi;
We are running Morningstar PPP (1.4b) and have been since '93 on a
worldblazer with no problems. I recently upgraded to a Microcom 28.8
modem. We leave our connection up 24 hours, 7 Days a week. The problem is
now I am getting Panics and Crashes when I turn on our minimal news feed.
I am wondering what's up or if anyone has seen this and can give guidence.
It seems entirely load/volume related. I am running this on an older less
used sun (IPX with ross cpu upgrade and 48MB, sunos 4.1.2) and I have the
serial port configured for 38400 but like I said it has been working fine
for years it seems to be speed related. Here is the info:
Port settings:
ttya "/usr/etc/getty std.38400" dialup on
ttya-mode=38400,8,n,1,-
ttya-rts-dtr-off=false
ttya-ignore-cd=false
Devices:
WORLDBLAZER cua0 38400 crtscts
/var/adm/messages::
Apr 8 06:29:41 eng1 vmunix: BAD TRAP
Apr 8 06:29:41 eng1 vmunix: pid 104, `pppd': Data fault
Apr 8 06:29:41 eng1 vmunix: kernel read fault at addr=0xff65f000, pme=0x0
Apr 8 06:29:41 eng1 vmunix: Sync Error Reg 80<INVALID>
Apr 8 06:29:41 eng1 vmunix: pc=0xf80cef14, sp=0xf836fa98, psr=0x900007c6,
context=0x7
Apr 8 06:29:41 eng1 vmunix: g1-g7: a00, f813a400, fd87e4c0, 1, f8370000, 0, 0
Apr 8 06:29:41 eng1 vmunix: Begin traceback... sp = f836fa98
Apr 8 06:29:41 eng1 vmunix: Called from f8059fb4, fp=f836faf8, args=ff65f000
fd8c8c20 4 66205269 6620 52690000
Apr 8 06:29:41 eng1 vmunix: Called from f80117dc, fp=f836fb58, args=ff655500
0 5dc fd8c8826 fd868c18 400
Apr 8 06:29:41 eng1 vmunix: Called from f80114a4, fp=f836fbb8, args=e ff655500 fd868c18 80fac 5ec 5dcApr 8 06:29:41 eng1 vmunix: Called from ff032a34,
fp=f836fc18, args=ff033528 ff655500 ff032fc4 fd8ea0f0 ff032c00 9704
Apr 8 06:29:41 eng1 vmunix: Called from ff032d80, fp=f836fc78, args=ff655500
ff033528 f836feac 908000e3 5dc 0
Apr 8 06:29:41 eng1 vmunix: Called from f8079370, fp=f836fce0, args=3b00
ff655500 0 ff033528 316913fc 3b00
Apr 8 06:29:41 eng1 vmunix: Called from f8063974, fp=f836fd68, args=3b00
f836feac 1 0 938 5dc
Apr 8 06:29:41 eng1 vmunix: Called from f804ec88, fp=f836fdc8, args=f827c354
1 f836feac 5dc fd85da94 0
Apr 8 06:29:41 eng1 vmunix: Called from f804dd2c, fp=f836fe40, args=f827c354
f836feac f836fea4 5dc 5dc f836feac
Apr 8 06:29:41 eng1 vmunix: Called from f80ee4c0, fp=f836fec0, args=f836ffe0
20 f811f8f8 f811f918 f8370000 f811f918
Apr 8 06:29:41 eng1 vmunix: Called from f8005a54, fp=f836ff58, args=f8370000 f836ffb4 f836ffe0 f8370000 f8370000 f836ffb4
Apr 8 06:29:41 eng1 vmunix: Called from 16a5c, fp=f7ffdc70, args=5 809d0 5dc 20 4000 3000
Apr 8 06:29:41 eng1 vmunix: End traceback...
----------------------------------------------
*************** showrev version 1.15 *****************
* Patch:
100103-11 - Set Permissions on system files. (05/05/93 -JDM)
100173-10 - NFS jumbo patch. (05/24/93 -JDM)
100448-01 - OW3 loadmodule security fix. (05/24/93 -JDM)
100444-58 - OW3.0: OW3.0 Server Patch (01/09/96 -JDM)
100512-04 - libXt CTE Jumbo Patch.(eng1) (01/09/96 -JDM)
100544-02 - Mailtool jumbo patch.(eng1) (05/24/93 -JDM)
100573-04 - libXmu Jumbo Patch.(eng1) (05/05/93 -JDM)
8.7.1 - New Sendmail (07/13/95 -JDM,NPM)
*******************************************************
Any suggestions or help appreciated!
-Jim
-----------------------------------------------------------------------------
Jim Murff, NIS (murff@nicimg.com) Voice # (619)635-8678
Nicolet Imaging Systems Inc., San Diego, CA. Corp # (619)695-6661
Senior Software Engineer/System Admin. Fax # (619)695-9902
-----------------------------------------------------------------------------
----- End Included Message -----
----- Begin Included Message -----
From: Kevin.Sheehan@uniq.com.au (Kevin Sheehan {Consulting Poster Child})
1) if you give us a symbolic traceback (run under kadb, do a $c when
it panics) it helps a lot more than random numbers.
2) given that you are on 4.1.2, my guess is the higher load just tickles
a STREAMS/zs bug that you hadn't seen before.
I'd take a look at the symbolic traceback and see what it suggests. It
may suggest upgrading :-)
l & h,
kev
----- End Included Message -----
----- Begin Included Message -----
From: Andrew J Cole <A.J.Cole@cbl.leeds.ac.uk>
Jim,
The error looks like the "getsockopt RESET trap" error. I have (out of
the blue) started to get BAD TRAPS after upgrading from Sendmail-8.7.4
to Sendmail-8.7.5. It seems that 8.7.5 tickles this particular problem
and several other people have noticed it. It also seems reasonable that
PPP might tickle it....
Mar 13 10:00:15 cblslca vmunix: BAD TRAP
Mar 13 10:00:15 cblslca vmunix: pid 20624, `sendmail': Data fault
Mar 13 10:00:15 cblslca vmunix: kernel read fault at addr=0x20, pme=0x0
Mar 13 10:00:15 cblslca vmunix: Sync Error Reg 80<INVALID>
Its difficult to say, for certain, that the patch cured the problem since
I was only getting 1-2 of these a week upto Mar 13 and none since I patched
the kernel.
Here are the patch numbers - I guess you will want 100804-03 since the
other one seems to be 4.1.3_U1 only.
Andrew
--------------------------------------------------------------------------------
Patch-ID# 100804-03
Keywords: getsockopt RESET trap mbuf leak bus panic TCP
Synopsis: SunOS 4.1.1,4.1.2,4.1.3: TCP socket and reset problems
Date: Jan/27/94
--------------------------------------------------------------------------------
Patch-ID# 101790-01
Keywords: getsockopt RESET trap mbuf leak bus panic TCP
Synopsis: SunOS 4.1.3_U1: TCP socket and reset problems
Date: Sep/12/94
--------------------------------------------------------------------------------
----- End Included Message -----
----- Begin Included Message -----
From: dhaut@level1.com (Dave Haut)
Jim,
You may want to try installing the tty jumbo patch for 4.1.2
( Patch 100513 ) ...
Hope this helps ...
---------------------------------
_ /| Dave Haut
\,o.O' Sys Admin
=(___)= Level One Communications
U dhaut@level1.com
(916) 854-4328
---------------------------------
----- End Included Message -----
----- Begin Included Message -----
From: pag@gpf.scg.boulder.co.us (Peter Gross)
You didn't say what version of the OS you were running -- SunOS 4.1.x or
Solaris 2.x. I have no experience with Solaris 2.x suns, but plenty with
SunOS 4.1.x. If that is your system you may need the Jumbo Tty Patch (it
was 100513-04 when I installed it; there seems to be a newer version
101621-03).
----- End Included Message -----
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:58 CDT