Re: [Openmosix-devel] kernel BUG with 2.4.27-om20041102-tab on SMP when doing heavy I/O

 moving to openmosix-general, as this message is more about asking for help
with ecc ram and stuff...

On Mon, Oct 03, 2005 at 07:10:41PM -0300, Peter Cordes wrote:
> On Sat, Oct 01, 2005 at 03:57:09AM +0200, Moshe Bar wrote:
> > Hard to think of where the bug (if any) might be.
> 
>  Yeah, I feel like I'm fumbling in the dark with unstable kernels, and
> nothing seems likely enough to merit seriously trying to do any hacking :(
> esp. since I'm still busy wrapping my head around cfengine2 and fai to keep
> my clusters manageable.  At least I'm confident the hardware is ok.  I've
> run memtester a lot while the system is up and running (and memtest86 for a
> few hours during some reboot downtime), so I don't think there's a problem
> there.

 On the other hand, I just had memtester find some problems.  It mlock()s
some pages and tests memory on a running system, so if something in the
kernel is corrupting memory in strange locations, it could be messing with
memtester's memory.  My machine has ECC RAM, and ECC is turned on in the
BIOS.  Does anyone know how to get Linux to report ECC events?  (On a Tyan
S2882, with an AMD chipset)?

 The (trimmed down) memtest output is below.  Note that the errors are
tightly clustered in time and in address location.  On a machine at home
which does seem to have maybe flaky hardware, I see similar clustering.
This makes me worried that I need to underclock my RAM or something for it
to be able to run error-free!  (At home, anyway (Athlon64 with a single 1GB
DDR400 DIMM from OCZ, running linux 2.6.12.6)...  I don't know about here
at work with ECC RAM and openMosix...)

nyneve:~# nice memtest 2000M
memtest v. 2.93.1
(C) 2000 Charles Cazabon <memtest-ltkGCvdZhqrnZt/[email protected]>
Original v.1 (C) 1999 Simon Kirby <[email protected]> <[email protected]>

Current limits:
  RLIMIT_RSS  0xffffffff
  RLIMIT_VMEM 0xffffffff
Raising limits...
Allocated 2097152000 bytes...trying mlock...success.  Starting tests...

Testing 2097147904 bytes at 0x40158000 (4088 bytes lost to page alignment).

...
Run    4:
  Test  1:         Stuck Address:  Testing...Passed.
  Test  2:          Random value:  Setting...Testing...Passed.
  Test  3:        XOR comparison:  Setting...Testing...Passed.
  Test  4:        SUB comparison:  Setting...Testing...Passed.
  Test  5:        MUL comparison:  Setting...Testing...Passed.
  Test  6:        DIV comparison:  Setting...Testing...Passed.
  Test  7:         OR comparison:  Setting...Testing...Passed.
  Test  8:        AND comparison:  Setting...Testing...Passed.
  Test  9:  Sequential Increment:  Setting...Testing...Passed.
  Test 10:            Solid Bits:  Testing...Passed.
  Test 11:      Block Sequential:  Testing...Passed.
  Test 12:          Checkerboard:  Testing...Passed.
  Test 13:            Bit Spread:  Testing...Passed.
  Test 14:              Bit Flip:  Testing...Passed.
  Test 15:          Walking Ones:  Testing...Passed.
  Test 16:        Walking Zeroes:  Testing...Passed.
Run    4 completed in 1858 seconds (0 tests showed errors).
Run    5:
...
  Test 11:      Block Sequential:  Testing...  87
FAILURE: 0x57575757 != 0x404bb080 at offset 0x0a481200.
Skipping to next test...
  Test 12:          Checkerboard:  Testing...Passed.
  Test 13:            Bit Spread:  Testing...Passed.
  Test 14:              Bit Flip:  Testing...Passed.
  Test 15:          Walking Ones:  Testing...Passed.
  Test 16:        Walking Zeroes:  Testing...Passed.
Run    5 completed in 1528 seconds (1 tests showed errors).
...
Run   15:
...
  Test 14:              Bit Flip:  Testing...Passed.
  Test 15:          Walking Ones:  Testing...   9
FAILURE: 0x00000200 != 0x20202020 at offset 0x0a481200.
Skipping to next test...
  Test 16:        Walking Zeroes:  Testing...  15
FAILURE: 0xffff7fff != 0x3532380a at offset 0x0a48120d.
Skipping to next test...
Run   15 completed in 1608 seconds (2 tests showed errors).
Run   16:
  Test  1:         Stuck Address:  Testing...   9
FAILURE: possible bad address line at offset 0x19e8112b.
Skipping to next test...
  Test  2:          Random value:  Setting...Testing...Passed.
  Test  3:        XOR comparison:  Setting...Testing...
FAILURE: 0x27067800 != 0x380a7800 at offset 0x0a4813ea.
Skipping to next test...
  Test  4:        SUB comparison:  Setting...Testing...
FAILURE: 0x2b911a6e != 0x3c951a6e at offset 0x0a4813ea.
Skipping to next test...
  Test  5:        MUL comparison:  Setting...Testing...
FAILURE: 0xcbb4bad4 != 0xdd4cbad4 at offset 0x0a4813ea.
Skipping to next test...
  Test  6:        DIV comparison:  Setting...Testing...
FAILURE: 0x00000001 != 0x00000000 at offset 0x0a481433.
Skipping to next test...
  Test  7:         OR comparison:  Setting...Testing...
FAILURE: 0xfb3d199f != 0xfb3d199e at offset 0x0a481433.
Skipping to next test...
  Test  8:        AND comparison:  Setting...Testing...Passed.
  Test  9:  Sequential Increment:  Setting...Testing...
FAILURE: 0x073fc6ea != 0x380ac6ea at offset 0x0a4814a9.
Skipping to next test...
  Test 10:            Solid Bits:  Testing...   9
FAILURE: 0x00000000 != 0x380a0000 at offset 0x0a481568.
Skipping to next test...
...
Run   16 completed in 1728 seconds (8 tests showed errors).
...
Run   43 completed in 1863 seconds (0 tests showed errors).
Run   44:
...
  Test 11:      Block Sequential:  Testing...  33
Received signal 2 (Interrupt)
munlock'ed memory.
43 runs completed.  11 errors detected.  Total runtime:  79568 seconds.

Exiting...



-- 
#define X(x,y) x##y
Peter Cordes ;  e-mail: X(peter@cor , des.ca)

"The gods confound the man who first found out how to distinguish the hours!
 Confound him, too, who in this place set up a sundial, to cut and hack
 my day so wretchedly into small pieces!" -- Plautus, 200 BC


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl