Unix server Panic: Kernel Mode Trap Errors?

Unix server Panic: Kernel Mode Trap Errors?

Post by Kiefer's Dre » Sun, 26 Oct 1997 04:00:00



Greetings,

We are having an ongoing problem with Panic KMT errors that crash our
Unix Server. We are using a software product called PC Connect to
establish telnet sessions to the Unix server across our Novell/Win95
office LAN.
The Unix Server is running Sunsoft Interactive Unix.
Most of the time this works fine but we are starting to get an
alarming occurance rate of Panic server crashes recently. We have a
shell inplace that bring's the server back up on its own but we are
trying to determine the cause of the problem.

After a system crash I would execute the following command to analyse
the crash.
wwpsams# crash -d /dev/swap
dumpfile = /dev/swap, namelist = /unix, outfile = stdout

trace command
results in...
STACK TRACE FOR PROCESS 65:
STKADDR   FRAMEPTR  FUNCTION  POSSIBLE ARGUMENTS
e0000d2c  e0000d40  sysdump  (8)
e0000d48  e0000d4c  panic    (d00c0059,e0000d68)
e0000d54  e0000d5c  cmn_err  (3,d00c0cc1,e,8c48300)
e0000d64  e0000d80  k_trap   (e0000d8c)
Trap e    e0000d8c  cmntrap  from d0086f95 in getcb
  ax: 8c48300 cx:    2414 dx:       8 bx:       0 fl:    10006 ds: 160
fs:   0
  sp:e0000dbc bp:e0000ddc si: 8c48300 di:      18 err:c1f50000 es: 160
gs:   0
e0000d94  e0000ddc  getcb    (18,d00c13d8,d0124408,d0119ab4)
e0000de4  e0000e00  wtcleanu (d0124408,d00c13d8,e0003537,d0119ab4)
e0000e08  e0000e40  vwxioctl (3537,620c,0,3)
e0000e48  e0000e60  s5ioctl  (d0119ab4,620c,0,3)
e0000e68  e0000e80  ioctl    (0,0,64,4016bc)
e0000e88  e0000eac  systrap  (e0000eb8)
          e0000eb8  sys_call from 0000a4b8
  ax:      36 cx:       0 dx:    aa86 bx:       0 fl:      296 ds:  1f
fs:   0
  sp:e0000ee8 bp:7ffff624 si:       2 di:      64 err:       7 es:  1f
gs:   0

stat command
results in...
system name:    wwpsams
release:        3.2
node name:      wwpsams
version:        2
machine name:   i386
time of crash:  Thu Oct 23 14:16:32 1997
age of system:  1 day, 23 hr., 34 min.
panicstr:       Kernel mode trap. Type 0x%x

and p command for process
results in...
  65 p  5731     1  5731  1794  64   0          wtm            load
  66 r  5912   565     0     0  96  41 00000000 amodac3.lm     load
  69 s  5734 19449  5734   301  30   0 d00d9520 sh             load
  70 s  5744  5734  5734   301  28   0 d0123ff8 amore.lm       load
  71 s 26066   284 26066     0  26   0 d010924c telnetd        load
  72 s 26067 26066 26067   678  28   0 d0124ef0 wtm            load
  73 s 26093 26067 26093   678  28   0 d0124c30 ddeserv.lm     load
  75 s  5938   503     0     0  39   0 e0000000 sleep          load
  77 r  5917  5912     0     0  20   7 00000000 amodac1.lm     load
nwak
  79 s  5914   565     0     0  10   1 d010c6bc amodac3.lm     load
nwak
  80 r  5915   565     0     0  81  11 00000000 amodac3.lm     load
  81 r  5916   565     0     0  79   6 00000000 amodac3.lm     load
  82 z  5926  5915     0     0  77   2 00000000 zombie         load
  83 s  5919  5912     0     0  27   0 d00d217e amodac2.lm     load
  85 z  5921  5916     0     0  10   0 00000000 zombie         load
  86 z  5923  5916     0     0  76   0 00000000 zombie         load
  89 s  5929  5915     0     0  27   0 d00d211e amodac2.lm     load
  92 z  5932  5914     0     0  76   0 00000000 zombie         load
  93 z  5933  5914     0     0  76   0 00000000 zombie         load

I am told that the stack trace is for the PID that caused the crash.
listing out the running processes at the chash shows that 65 was
a wtm or a windows terminal load attempt.
65 p  5731     1  5731  1794  64   0          wtm            load

This is always what I see after a crash. The 1794 user id is usually
the culprit. This connection is being made through our WAN from an
ajoining facility. Once the system is back up and running the zombie's
are gone but I am wondering what relationship they might have with the
crashes. I usually always see at least one zombie process in the
dumpfile analysis.

I am wondering if anyone can offer any advice as to what to try next
to eliminate this problem from occuring in the future. Should we be
locking at the network card in the server or some sort of hardware
issue related to this problem?

Any help would be much appreciated.
Thanks in Advance
Phil,
--

Please remove NOSPAM from email when replying,

"We live to survive our paradoxes..."
G. Downie. http://www.thehip.com

 
 
 

1. Kernel mode error - PANIC: k-trap

We have a customer site where about every 3-4 weeks we get the following
error which shuts the system down:

Unexpected trap in kernel modem
(registers)

PANIC: k-trap - kernel modem trap 0x0000000E
trying to dump 4000 pages to dumpdev hd 1/41 at block 0

This PANIC does not get written to the /usr/adm/messages file

No other indications of a problem.  They turn the machine off and back off
(after a brief pause) and then it will run fine for another month....

This is a 486dx50 Mhz machine, with 16mb of memory, Adaptec 1542CF
controller, SCSI 670mb HD, Wangtek 250mb tape, dual floppy, Digiboard PCx
16 port multiport controller, running SCO UNIX 3.2v4.2
The software they run is RealWorld v6.5 and Crossroads Dispatching software.

I am at a loss to see what is going on here.  Nothing was found by looking
in the IT scripts.

Thanks for any comments.

--

#.. CCS Enterprises, Inc.    ..  |  ..   Office: (410) 820-4670          ..#
#.. P.O. Drawer 1690         ..  |  ..      FAX: (410) 476-5261          ..#
#.. Easton, Maryland 21601 ......|......... BBS: (410) 476-5098 ...........#

2. NFS Bug In UnixWare 2.03

3. Interactiv Unix - Panic Kernel Mode trap

4. Linux on Quaderno 33 ????

5. PANIC: K-TRAP KERNAL MODE TRAP TYPE 0x0000006

6. Help with Mini-Linux prompt!!

7. "Unexpected trap in kernel mode" panic.

8. SA-286 Test Questions/Review Url?

9. SCO Panic - Unexpected trap in kernel mode

10. PANIC: k_trap kernel mode trap type 0x0000000E

11. Help with Panic Trap Kernel ERROR:

12. "Panic: K_trap - Kernal mode trap type 0x00000000" HELP

13. ESIX 4.0.3A Panic Trap Mode 0xD, Please help trace.