Hi

I am getting this error messages on my HP DL580 running RHEL3/U4. Server crashed. Apologies if I am posting this in the wrong thread.

Any ideas - not sure what the NVidia message is or is it a HW problem with cpu5

Thanks

Mick

Nov 10 18:57:34 testdb2 kernel: Page has mapping still set. This is a serious situation. However if you
Nov 10 18:57:35 testdb2 kernel: are using the NVidia binary only module please report this bug to
Nov 10 18:57:35 testdb2 kernel: NVidia and not to the linux kernel mailinglist.
Nov 10 18:57:35 testdb2 kernel: ------------[ cut here ]------------
Nov 10 18:57:35 testdb2 kernel: kernel BUG at page_alloc.c:225!
Nov 10 18:57:35 testdb2 kernel: invalid operand: 0000
Nov 10 18:57:35 testdb2 kernel: nfs lockd sunrpc hangcheck-timer usbserial lp parport autofs4 ocfs audit tg3 floppy sg microc
ode keybdev mousedev hid input ehci-hcd usb-uhci usbcore ext3 jbd
Nov 10 18:57:35 testdb2 kernel: CPU: 5
Nov 10 18:57:35 testdb2 kernel: EIP: 0060:[<c0158800>] Not tainted
Nov 10 18:57:35 testdb2 kernel: EFLAGS: 00210286
Nov 10 18:57:35 testdb2 kernel:
Nov 10 18:57:35 testdb2 kernel: EIP is at __free_pages_ok [kernel] 0x3e0 (2.4.21-27.ELsmp/i686)
Nov 10 18:57:35 testdb2 kernel: eax: 00000033 ebx: c30bd15c ecx: 00000001 edx: c0381e98
Nov 10 18:57:35 testdb2 kernel: esi: c30bd15c edi: 001b6000 ebp: 00000000 esp: f4c29e10
Nov 10 18:57:35 testdb2 kernel: ds: 0068 es: 0068 ss: 0068
Nov 10 18:57:35 testdb2 kernel: Process oracle (pid: 6987, stackpage=f4c29000)
Nov 10 18:57:35 testdb2 kernel: Stack: c02bd994 f7f16008 000000fc 00200286 0000001e c3067b30 00000000 00200286
Nov 10 18:57:35 testdb2 kernel: ca4a9afc ee573200 fff0edb0 00000000 001b6000 8baf4047 c30bd15c 001b6000
Nov 10 18:57:35 testdb2 kernel: fff0edb0 c013ef62 c30bd15c 8baf4047 00000000 c0145187 f4cf9d7c 777b6000
Nov 10 18:57:35 testdb2 kernel: Call Trace: [<c013ef62>] __free_pte [kernel] 0x52 (0xf4c29e54)
Nov 10 18:57:35 testdb2 kernel: [<c0145187>] zap_pte_range [kernel] 0x2a7 (0xf4c29e64)
Nov 10 18:57:35 testdb2 kernel: [<c013ff8b>] zap_page_range [kernel] 0x14b (0xf4c29ec8)
Nov 10 18:57:35 testdb2 kernel: [<c017d740>] dput [kernel] 0x30 (0xf4c29edc)
Nov 10 18:57:35 testdb2 kernel: [<c0146c3d>] do_munmap [kernel] 0x25d (0xf4c29f1c)
Nov 10 18:57:35 testdb2 kernel: [<c0172765>] path_release [kernel] 0x15 (0xf4c29f34)
Nov 10 18:57:35 testdb2 kernel: [<c01af517>] sys_shmdt [kernel] 0xa7 (0xf4c29f68)
Nov 10 18:57:35 testdb2 kernel: [<c0113aac>] sys_ipc [kernel] 0x22c (0xf4c29f90)
Nov 10 18:57:35 testdb2 kernel: [<c01633e2>] sys_open [kernel] 0xa2 (0xf4c29fa8)
Nov 10 18:57:35 testdb2 kernel:
Nov 10 18:57:35 testdb2 kernel: Code: 0f 0b e1 00 17 d2 2b c0 e9 6c fc ff ff 9c 5a fa f0 fe 0d ec
Nov 10 18:57:35 testdb2 kernel:
Nov 10 18:57:35 testdb2 kernel: Kernel panic: Fatal exception
Nov 10 18:57:35 testdb2 kernel: kernel BUG in header file at line 56
Nov 10 18:57:35 testdb2 kernel: ------------[ cut here ]------------
Nov 10 18:57:35 testdb2 kernel: kernel BUG at panic.c:311!
Nov 10 18:57:35 testdb2 kernel: invalid operand: 0000
Nov 10 18:57:35 testdb2 kernel: nfs lockd sunrpc hangcheck-timer usbserial lp parport autofs4 ocfs audit tg3 floppy sg microc
ode keybdev mousedev hid input ehci-hcd usb-uhci usbcore ext3 jbd

Did you just recently upgrade to U4 or the kernel perhaps? Has this always occured or did it start crashing after any upgrades? Seems to me like it could be a kernel issue.

What version and what version of Oracle are you running?

No OS or Kernel upgrade done on this server and it is Oracle 9.2.0

Any other changes, additions, new data loaded into Oracle? Did this always happen or recently? Does it occur rather quickly or over time?

No changes were made to the server - application testing against one the DBs may have being happening. I just want to know is this an HW issue on that CPU

Could be. Best way to troubleshoot hardware issues is to remove the culprit, see if it occurs without it

It doesn't look like you have the nvidia module loaded, so that's not an issue.

It looks like the process that caused the panic was your oracle DB. It seems you're running an SMP kernel and have multiple processors. Is Oracle 9.2.0 thread-safe for x86 platforms? It might be a bug in either Oracle or that kernel that there is some kind of race condition freeing pointers on a multi-CPU system.