Forum Discussion

Roman_Kuznetsov's avatar
Roman_Kuznetsov
Level 3
12 years ago

Solaris 11.1 VXVM 6.0.1 'df' cause a panic

Environment:

System Configuration: HP ProLiant BL480c G1

Oracle Solaris 11.1 X86

panic string:  

BAD TRAP: type=e (#pf Page fault) rp=fffffffc816fdb90 addr=0 occurred in module "unix" due to a NULL pointer dereference

Veritas INFO: 

PKGINST:  VRTSvxvm
      NAME:  Binaries for VERITAS Volume Manager by Symantec
  CATEGORY:  system
      ARCH:  i386
   VERSION:  6.0.100.000,REV=08.01.2012.08.52

Stack:

genunix: [ID 655072 kern.notice] fffffffc816fdab0 unix:die+105 ()
genunix: [ID 655072 kern.notice] fffffffc816fdb80 unix:trap+153e ()
genunix: [ID 655072 kern.notice] fffffffc816fdb90 unix:cmntrap+e6 ()
genunix: [ID 655072 kern.notice] fffffffc816fdca0 unix:strncpy+1c ()
genunix: [ID 655072 kern.notice] fffffffc816fdcd0 odm:odmstatvfs+90 ()
genunix: [ID 655072 kern.notice] fffffffc816fdcf0 genunix:fsop_statfs+1a ()
genunix: [ID 655072 kern.notice] fffffffc816fde70 genunix:cstatvfs64_32+42 ()
genunix: [ID 655072 kern.notice] fffffffc816fdec0 genunix:statvfs64_32+69 ()
genunix: [ID 655072 kern.notice] fffffffc816fdf10 unix:brand_sys_sysenter+1dc ()

Messages:


unix: [ID 839527 kern.notice] df:
unix: [ID 753105 kern.notice] #pf Page fault
unix: [ID 532287 kern.notice] Bad kernel fault at addr=0x0
unix: [ID 243837 kern.notice] pid=3965, pc=0xfffffffffb893ff8, sp=0xfffffffc816fdc88, eflags=0x10206
unix: [ID 211416 kern.notice] cr0: 80050033<pg,wp,ne,et,mp,pe> cr4: 6f8<xmme,fxsr,pge,mce,pae,pse,de>
unix: [ID 624947 kern.notice] cr2: 0
unix: [ID 625075 kern.notice] cr3: 59f0a2000
unix: [ID 625715 kern.notice] cr8: c
unix: [ID 100000 kern.notice]
unix: [ID 592667 kern.notice]       rdi: fffffffc816fdd48        rsi:                0             rdx:                f
unix: [ID 592667 kern.notice]       rcx:                1          r8:              e80             r9:                0
unix: [ID 592667 kern.notice]       rax: fffffffc816fdd48       rbx:         fefa3430          rbp: fffffffc816fdca0
unix: [ID 592667 kern.notice]       r10: fffffffffb856d00         r11:                0            r12: fffffffc816fdd00
unix: [ID 592667 kern.notice]       r13: ffffc10012176880    r14:                0            r15: ffffc1002bb09480
unix: [ID 592667 kern.notice]       fsb:                0           gsb: ffffc1000eac8000     ds:               4b
unix: [ID 592667 kern.notice]        es:               4b          fs:                0               gs:              1c3
unix: [ID 592667 kern.notice]       trp:                e            err:                0             rip: fffffffffb893ff8
unix: [ID 592667 kern.notice]        cs:               30           rfl:            10206            rsp: fffffffc816fdc88
unix: [ID 266532 kern.notice]        ss:               38

In preced log of panic I see "odm:odmstatvfs+90". I think this is root of panic, but due in lack of scat and mdb knowlage, I am cannot to investigate this module. When I delete VXVM, there is no panic when I issue the 'df'.

If I can provide more information about this case, please let me know. For now I dont know what additional info to provide.

Core dump is about of 400 MB, which is more than I can attach to this message.

business continuity
Solaris
Storage Foundation
Troubleshooting
  • mikebounds's avatar
    mikebounds
    12 years ago

    If you think issue is odm you could:

    1. umounted /dev/odm if it umounted
    2. Stop odm is it is running (it shows in output of gabconfig -a if it is running as port d I think)
    3. Unload odm from kernel ("modinfo | grep odm" and then modunload the id given)

    Then see if df still panics the system.

    ODM is an optional component used to improve the access to Oracel disk files

    Mike

13 Replies

Sort By
  • mikebounds's avatar
    mikebounds
    Level 6
    12 years ago

    If you think issue is odm you could:

    1. umounted /dev/odm if it umounted
    2. Stop odm is it is running (it shows in output of gabconfig -a if it is running as port d I think)
    3. Unload odm from kernel ("modinfo | grep odm" and then modunload the id given)

    Then see if df still panics the system.

    ODM is an optional component used to improve the access to Oracel disk files

    Mike

  • Roman_Kuznetsov's avatar
    Roman_Kuznetsov
    Level 3
    12 years ago

    So nice! We got it!

    I know that I have no support of glorious company Veritas, but this can be a bug. It would be nice to bug report or some. How I can to do this, and whether exist the reason for it?

    Thank you, Mike!

  • mikebounds's avatar
    mikebounds
    Level 6
    12 years ago

    I will report this, so you may be contacted on this forum for more details.

    Mike

  • TonyGriffiths's avatar
    TonyGriffiths
    Level 6
    12 years ago

    Hi,

    We are looking to claim support of Solaris 11.1 with SFHA 6.0.3 due soon. I will post an update when 6.0.3 has been released.

    cheers

    tony

  • mikebounds's avatar
    mikebounds
    Level 6
    12 years ago

    Brad, you originally wrote:

    I don't think 6.0.3 fixes the issue.� I have a Solaris 11.1(SPARC) install that would panic when 'df' was issued using 6.0.1.� I just installed 6.0.3 and it still panics with 'df

    But then 3 hours later you changed your comment to "...".  Does this mean that 6.0.3 does resolve the df panic which is why you removed the comment?

    Mike

  • BradMc's avatar
    BradMc
    Level 3
    12 years ago
    Sorry, I tried to delete the message quickly, as I saw there was need to keep 6.0.1 unconfigured before installing 6.0.3. I rebuilt the system, installed 6.0.1, then installed 6.0.3, and it worked. One note, though, using the text installer, you can't tell the installer to keep the 6.0.1 install unconfigured. It doesn't ask the question as in previous versions. However, by answering "No" to install license key, it will stay unconfigured.