supervisor crashes xserve with malloc errors

[westernx]

. . . me again

About once a day our supervisor is crashing our main os x server.  I checked the logs and this is what I found in the supe.log ;

Code: [Select]

*** set a breakpoint in malloc_error_break to debug
terminate called after throwing an instance of 'std::bad_alloc'
  what():  St9bad_alloc
supervisor(2254) malloc: *** mmap(size=1595867136) failed (error code=12)
*** error: can't allocate region
*** set a breakpoint in malloc_error_break to debug
terminate called after throwing an instance of 'std::bad_alloc'
  what():  St9bad_alloc
. . . seems like some sort of memory error.  Could this be generated from bad memory modules in the server?  

R

[westernx]

. . . actually looks like RAM is o.k

Code: [Select]

Diagnostics:

    Power On Self-Test:

      Last Run: 7/27/10 2:58 PM
      Result: Passed

---------------------------------------------

      BRANCH 0 CHANNEL 0/DIMM 1:

          Size: 1 GB
          Type: DDR2 FB-DIMM
          Speed: 667 MHz
          Status: OK
          Manufacturer: 0x80AD
          Part Number: 0x48594D5031313241373243503844332D5335
          Serial Number: 0x6777131A
--
        BRANCH 0 CHANNEL 1/DIMM 2:

          Size: 1 GB
          Type: DDR2 FB-DIMM
          Speed: 667 MHz
          Status: OK
          Manufacturer: 0x80AD
          Part Number: 0x48594D5031313241373243503844332D5335
          Serial Number: 0x67771819
--
        BRANCH 1 CHANNEL 0/DIMM 3:

          Size: 1 GB
          Type: DDR2 FB-DIMM
          Speed: 667 MHz
          Status: OK
          Manufacturer: 0x80AD
          Part Number: 0x48594D5035313241373243503844332D5335
          Serial Number: 0x61792306
--
        BRANCH 1 CHANNEL 1/DIMM 4:

          Size: 1 GB
          Type: DDR2 FB-DIMM
          Speed: 667 MHz
          Status: OK
          Manufacturer: 0x80AD
          Part Number: 0x48594D5035313241373243503844332D5335
          Serial Number: 0x66957911
--
R

[shinya]

Hi R,

We've not seen this happen to others.
What version of the OS is this?

As you suspect, I also think it's likely a hardware issue.

The MacOS memory test is known to often "pass" bad memory.
Having said that, you could first try running the memory test in single-user
mode for a little more thorough RAM test. If it still doesn't catch an issue,
you may want to try a third-party memory testing app, such as "memtest"
http://www.memtestosx.org/ (I just checked, but they don't have a OSX 10.6
version, though).

Let us know how these tests go.

-shinya.