PFN_LIST_CORRUPT BSOD on Proliant DL380G4

[Yurij Salmaniw]

Whenever I do file copies, whether xcopy off the network to another server, or tape backup onto an HP SDLT320 drive, I receive on a regular basis, usually after a while into the process, the BSOD message:

PFN_LIST_CORRUPT
0x0000004e (0x0000009a,0x000005659, 0x00000006, 0x00000002)

There is no event in the log file to indicate a problem, only after core minidump and reboot does the event log show the above message.

I have 6i used for internal mirror of two 36GB drives, and the LSI Ultra320 connectec to HP SDLT 320 drive. Using single gigabit port full duplex to connect to iSCSI device using MS iSCSI initiator v2.0

Based on what I've read about this error, it can be a driver issue and tracking of physical pages of memory corruption, or incorrect MDL handling; or bad memory.

I have included the minidump of the latest crash.

Researching this matter has indicated that it may be caused by bad memory ... I am looking for feedback before I do further low level diagnostics on the MB.

Any advice?

[John Marinov]

Hi Yurij,

The minidump says that the cause was memory corruption, so it's most likely a physical problem. Run the intensive diagnostics or swap out the memory (or swap it around).

First thing to check is RAM, which will be pretty easy to eliminate. Swap the memory around in the box, or better yet, swap it for known good memory. Don't forget that the memory slots can also be a problem -which is a mainboard change.

It looks like large data copies off the box give problems. Assuming all your drivers and firmware are up to spec, and that your system memory is good, the only thing left is the system main board. Specifically the system bus.

This is the result of a kernel debug analysis of your minidump.

FOLLOWUP_IP:
nt!MiBadRefCount+33
80888132 cc int 3
SYMBOL_STACK_INDEX: 1
FOLLOWUP_NAME: MachineOwner
SYMBOL_NAME: nt!MiBadRefCount+33
MODULE_NAME: nt
DEBUG_FLR_IMAGE_TIMESTAMP: 42435e60
STACK_COMMAND: kb
IMAGE_NAME: memory_corruption
FAILURE_BUCKET_ID: 0x4E_nt!MiBadRefCount+33
BUCKET_ID: 0x4E_nt!MiBadRefCount+33

Notice that KD blames "memory corruption".

Cheers,

[Yurij Salmaniw]

Thanks for the quick response. To be assured this is the problem I have uploaded a couple more minidumps to this url:

http://users.eastlink.ca/~salmageo/dmp.htm


Do these also show a memory problem?