PANIC - need assistance

PANIC - need assistance

Post by Michel P. Pouli » Wed, 08 Nov 2000 04:00:00



I'm having a problem w/ one of my clients servers.  Here's the lowdown:

First the error:
Unexpected trap in kernel mode:
cr0 0x8001003B     cr2  0x00000002     cr3 0x00002000     tlb  0x00000000
ss  0x0000BD50     uesp 0xF01E5DF0     efl 0x00010202     ipl  0x00000007
cs  0x00000158     eip  0x00000047     err 0x00000003     trap 0x0000000E
eax 0x00000002     ecx  0x00000000     edx 0xFFFFFFFF     ebx  0x00000100
esp 0xE0000C48     ebp  0xE0000C70     esi 0xF02ABD50     edi  0x00000000
ds  0x00000160     es   0x00000160     fs  0x00000000     gs   0x00000000
cpu 0x00000001

PANIC: k_trap - Kernel mode trap type 0x0000000E
Unable to freeze processor 2, proceeding...
Cannot dump 65407 pages to dumpdev hd (1/41): Space for only 64000 pages

Dump not completed

Now the specs:
uname -X
SCO=5.0.4

# custom
Apache Web Server (ver 1.3.12)
Compaq Extended Feature Supplement (ver 5.21a)                       ||
||       DigiWARE for RealPort Driver (ver 2.0.2)
 ||
||       NetTerminal Host software for SCO OpenServer (ver 1.1.0b)
 ||
||       Olympus Software, Inc. Olympus TuneUp (ver 6.0.0L)
 ||
||       SCO OpenServer Enterprise System (ver 5.0.4p)
 ||
||       SCO Symmetrical Multiprocessing (ver 1.1.0Eb)
 ||
||       SCO VisionFS (ver 3.00.917)
 ||
||       Release Supplement for SCO OpenServer Release 5.0.4 (ver rs504c)
 ||
||       Release Supplement for SCO OpenServer Release 5.0.4 (ver
rs.Unix504. ||
||       Software Manager Supplement (ver rs504c)
 ||
||       Year 2000 Supplement for SCO OpenServer 5.0.2 and 5.0.4 (ver
oss601a ||
||       oss601a Software Manager patch FOR OSR 5.0.4 ONLY (ver
oss601a.SoftM

# cat /var/adm/syslog
Nov  7 03:10:23 car-tech device    address      vec dma  comment
Nov  7 03:10:23 car-tech
-------------------------------------------------------
------------------------
Nov  7 03:10:24 car-tech %cpu      -              -   -  unit=1 family=6
type=Un
known
Nov  7 03:10:24 car-tech %cpuid    -              -   -  unit=1
vend=GenuineInte
l tfms=0:6:5:1
Nov  7 03:10:24 car-tech %fpu      -             13   -  unit=1
type=80387-compa
tible
Nov  7 03:10:24 car-tech %pci      0x0CF8-0x0CFF          -   -  am=1 sc=1
buses
=3
Nov  7 03:10:24 car-tech %serial   0x03F8-0x03FF          4   -  unit=0
type=Sta
ndard nports=1 fifo=yes
Nov  7 03:10:24 car-tech %console  -              -   -  unit=vga type=0 12
scre
ens=68k
Nov  7 03:10:24 car-tech %floppy   0x03F2-0x03F7          6   2  unit=0
type=135
ds18
Nov  7 03:10:24 car-tech %ida0     0x3000-0x301F         11   -  PCI slot=2
unit
s=1 Compaq IDA/SMART (rev 5.21a)
Nov  7 03:10:24 car-tech %adapter  0x01F0-0x01F7         14   -  type=IDE
ctlr=p
rimary   dvr=wd
Nov  7 03:10:23 car-tech ifor_pmd:
Nov  7 03:10:24 car-tech %adapter  0x2000-0x20FF          9   -  type=cha
ha=1 i
d=7 PCI slot=0 geo=B wid=16 Compaq Fast-SCSI-2 (rev 5.21a)
Nov  7 03:10:25 car-tech %adapter  0x2400-0x24FF         10   -  type=cha
ha=0 i
d=7 PCI slot=0 geo=B wid=16 Compaq Fast-SCSI-2 (rev 5.21a)
Nov  7 03:10:25 car-tech %csmu     -              -   -  Compaq SCSI-2
Managemen
t driver (rev 5.21a)
Nov  7 03:10:25 car-tech %cled     -              -   -  Compaq ProLiant
Storage
 System (rev 5.21a)
Nov  7 03:10:25 car-tech %pcibios  -              -   -  PCI BIOS (2.16),
Hardwa
re Mechanism 1
Nov  7 03:10:25 car-tech syslogd: re-initialise
Nov  7 03:10:25 car-tech syslogd: restart
Nov  7 03:10:25 car-tech %cpqw     -              -   -  Compaq Wellness
driver
(ver 5.21a)
Nov  7 03:10:25 car-tech %crom     -              -   -  Compaq rom driver
(rev
5.21a)
Nov  7 03:10:25 car-tech %cnet     0x2800-0x280F          5   -  unit=0
Nov  7 03:10:25 car-tech RealPort Driver 2.0.2
Nov  7 03:10:25 car-tech %OlyTune  -              -   -  Olympus Software,
Inc.
System TuneUp
Nov  7 03:10:25 car-tech %OlyTune  -              -   -  Olympus TuneUp
Release
6.0.0L
Nov  7 03:10:26 car-tech %cd-rom   -              -   -  type=IDE ctlr=pri
cfg=m
st dvr=Srom->wd
Nov  7 03:10:26 car-tech %tape     -              -   -  type=S ha=1 id=6
lun=0
bus=0 ht=cha
Nov  7 03:10:26 car-tech %disk     -              -   -  type=IDA0 unit=0
cyls=6
533 hds=255 secs=32
Nov  7 03:10:26 car-tech
Nov  7 03:10:26 car-tech NOTICE: TuneUp OK status1: 0405BF0F F00CDEF8
55F024F7 E
C83EC8B
Nov  7 03:10:26 car-tech
Nov  7 03:10:26 car-tech NOTICE: TuneUp OK status1: EBEC8B55 F0114D84
909090B9 8
BC03355
Nov  7 03:10:26 car-tech
Nov  7 03:10:26 car-tech NOTICE: TuneUp OK status1: EBEC8B55 F00158C0
909090F9 4
2B809EB
Nov  7 03:10:26 car-tech mem: total = 261628k, kernel = 209396k, user =
52232k
Nov  7 03:10:26 car-tech swapdev = 1/41, swplo = 0, nswap = 524290, swapmem
= 26
2144k
Nov  7 03:10:26 car-tech rootdev = 1/42, pipedev = 1/42, dumpdev = 1/41
Nov  7 03:10:26 car-tech kernel: Hz = 100, i/o bufs = 179176k (high bufs =
17380
4k)
Nov  7 03:10:26 car-tech
Nov  7 03:10:26 car-tech %cpu      -            255   -  unit=2 family=6
type=Un
known
Nov  7 03:10:26 car-tech %cpuid    -              -   -  unit=2
vend=GenuineInte
l tfms=0:6:5:1
Nov  7 03:10:26 car-tech %fpu      -              -   -  unit=2
type=80387-compa
tible

trap trigured by:
any port 23 manipulation, attempts to remove or replace Netterm software, or
update system w/ OSS469d

Most recent changes:
Apache installed a week ago and Netscape Fast Track removed. (all using
apache install process)
10 user license bump applied  a few days later w/ kernel re-link but did not
re-boot right away.

We have other similar systems w/ Apache or Netterm software, but this is the
only client we have w/ both software packages on a 5.0.4 system.

The day the server was re-booted, the error started.  All hardware looked at
and parts replaced due to poor power supply, but the problm returned after
restore. from tape.

Server running on Compaq Proliant 1600

Any ideas?

--
-=Michel P, Poulin=-
SCO Support - MicroMini Associates, Inc.

 
 
 

PANIC - need assistance

Post by Jim Bonne » Wed, 08 Nov 2000 04:00:00


Try installing oss469d, This fixed a problem with Unable to freeze
processors, although a Ktrap is usually a hardware thing and the
0x000000E means a page fault maybe you have bad RAM?

I'd still put on 469d though..

Quote:> PANIC: k_trap - Kernel mode trap type 0x0000000E
> Unable to freeze processor 2, proceeding...
> Cannot dump 65407 pages to dumpdev hd (1/41): Space for only 64000 pages

--
Jim Bonnet

 
 
 

PANIC - need assistance

Post by Tony Lawrenc » Wed, 08 Nov 2000 04:00:00



> I'm having a problem w/ one of my clients servers.  Here's the lowdown:
> PANIC: k_trap - Kernel mode trap type 0x0000000E
> The day the server was re-booted, the error started.  All hardware looked at
> and parts replaced due to poor power supply, but the problm returned after
> restore. from tape.

> Server running on Compaq Proliant 1600

> Any ideas?

Trap E is almost always hardware (
http://aplawrence.com/Unixart/trape.html ), though for an SMP
machine I've seen exactly this when the SMP supplement is not
installed.  However, you seem to have been running without it and
this just suddenly happened, so it looks like hardware.  As a
guess, I'd bet the nic.., though system memory is also a
possibility if it's really related to the license bump because
that does change some kernel tables which would rearrange things
in memory- have you tried booting unix.old to eliminate that
theory?

BTW, this post does NOT belong in comp.unix.sco.programmer
because it's not about programming, does not belong in
sco.opendesktop because it's not ODT and I don't think that's
valid anymore anyway, and it looks like you made up the rest of
the newsgroups you posted to- very, very bad form..

--

SCO/Linux articles, help, book reviews, tests,
job listings and more : http://www.pcunix.com

 
 
 

PANIC - need assistance

Post by Dirk Har » Wed, 08 Nov 2000 04:00:00



> Try installing oss469d, This fixed a problem with Unable to freeze
> processors, although a Ktrap is usually a hardware thing and the
> 0x000000E means a page fault maybe you have bad RAM?

> I'd still put on 469d though..

I think he said he tried that - there was a lot of stuff to read though.

        trap trigured by:
        any port 23 manipulation, attempts to remove or replace Netterm
software, or
        update system w/ OSS469d

RAM is a good idea though.  Try rotating the SIMMs/DIMMs/whatever to see
if that changes the situation.

And, in case you haven't, shutdown and *power off*, then reboot.

 
 
 

PANIC - need assistance

Post by Michel P. Pouli » Wed, 08 Nov 2000 04:00:00


I thought it was memory as well, but do not have a memory tester to prove
this theory.  Switching memory sticks around has its pros and cons, and may
change the address of the error if I'm lucky.

This machine has it's SMP license installed, and has had it since build-up.

When Apache is removed, it does appear to resolve the issue. This
installation has been runnig for over a year w/ Fasttrack Unfortunatly,
Fasttrack has some issues that cause problems w/ our applications that use
http get and put.

BTW,  I posted to any newsgroups that may be of use.  I've had Compaq on
site telling me it's a software issue, thus I need input from the
programmers and anyone one else w/ input.  I will admit that I misread
opendesktop for openserver though. Also the newgroups I posted to show up in
my list of available groups, I didn't make any of them up.

--
-=Michel P, Poulin=-
SCO Support - MicroMini Associates, Inc.

 
 
 

PANIC - need assistance

Post by Tony Lawrenc » Wed, 08 Nov 2000 04:00:00



> I thought it was memory as well, but do not have a memory tester to prove
> this theory.  Switching memory sticks around has its pros and cons, and may
> change the address of the error if I'm lucky.

Again- have you tried unix.old?  

Quote:

> This machine has it's SMP license installed, and has had it since build-up.

I still think you need the SMP supplement- I had occaasional
panics on an HP SMP until this was installed..

Quote:> BTW,  I posted to any newsgroups that may be of use.  I've had Compaq on
> site telling me it's a software issue, thus I need input from the
> programmers and anyone one else w/ input.

Well, the programmers group is mostly about writing software,
using the compilers, etc., and not about solving OS software
problems.  But you are in the right place now.

Quote:>I will admit that I misread
> opendesktop for openserver though. Also the newgroups I posted to show up in
> my list of available groups, I didn't make any of them up.

Hmm.. I don't see any of comp.unix.sco, sco.scounix, sco.unix ..
but it doesn't matter, this is where you want to be.

--

SCO/Linux articles, help, book reviews, tests,
job listings and more : http://www.pcunix.com

 
 
 

PANIC - need assistance

Post by Michel P. Pouli » Wed, 08 Nov 2000 04:00:00



Quote:

> Again- have you tried unix.old?

Unfortunatly, the system has been restored from tape before I could try
this.
Quote:

> I still think you need the SMP supplement- I had occaasional
> panics on an HP SMP until this was installed..

I'll look into this.

The tech on site seems to believe that this issue is somehow related to
Boundless netterm software changing telnetd (or another port 23 watching
utility) to allow for their special controll features, and Apache changing
it as well for security.  THe oddest part is that this doesn't cause a
problem untill a relink and re-boot is done, and Apache doesn't need a
relink to make it work.  I'm about to dig into their website to find out
what their install changes and compait that to boundless.

Thanks for the info thus far!

--
-=Michel P, Poulin=-
SCO Support - MicroMini Associates, Inc.