3 Replies Latest reply: Feb 1, 2013 2:48 AM by 908067 RSS

    reboot cause

      Hi guys,

      I have a Solaris 10 server that got rebooted...Vmcore file created..I am a newbie in this filed, but I tried to analyze the cause of reboot anyway. I used 2 tools to analyze the crash dump and got to 2 different conclusions, so I need help to clarify the real issue.

      Using mdb, looks like the panic was created by an Oracle-RAC related process (oprocd.bin):
                   cpu                1
                thread      30004243500
               message forced crash dump initiated at user request
                tstate       4400001605
                    g1                b
                    g2                0
                    g3          11f7a44
                    g4              6e0
                    g5         88000000
                    g6                0
                    g7      30004243500
                    o0          128f9f8
                    o1      2a100c619e8
                    o2                1
                    o3                0
                    o4 fffffffffffffff5
                    o5             1000
                    o6      2a100c610b1
                    o7          10701a8
                    pc          104abec
                   npc          104abf0
                     y                0
       30004243500::thread -p
                  ADDR             PROC              LWP             CRED
      0000030004243500      6001c0f5968      600200659e0      60010025e38
       6001c0f5968::ps -t
      S    PID   PPID   PGID    SID    UID      FLAGS             ADDR NAME
      R   4864   4707   1118   1118      0 0x4a004000 000006001c0f5968 oprocd.bin
              T     0x30004243500 <TS_ONPROC>
      Don't ask me how I ran te above commands, just used a similar info from web :)

      Using SCAT, looks like the crash was forced from keyboard (STOP+A):
      CAT(vmcore.2/10U)> analyze
      core file:      /var/crash/xxxxxx/vmcore.2
      user:           Super-User (root:0)
      release:        5.10 (64-bit)
      version:        Generic_147440-09
      machine:        sun4u
      node name:      xxxxxxxx
      domain:        xxxxx
      hw_provider:    Oracle Corporation
      system type:    SUNW,Sun-Fire-V240 (UltraSPARC-IIIi)
      hostid:         xxxxx
      dump_conflags:  0x10000 (DUMP_KERNEL) on /dev/dsk/c1t0d0s1(3.90G)
      time of crash:  Wed Jan 30 01:06:32 MST 2013
      age of system:  57 days 13 hours 34 minutes 22.61 seconds
      panic CPU:      1 (2 CPUs, 8G memory, 1 nodes)
      panic string:   forced crash dump initiated at user request
      ==== system appears to have been Stop-A'ed ====
      Of course, the output is longer; if relevant I'll paste here...

      Now, which source should I trust?

      Thank you!
        • 1. Re: reboot cause
          The sources are not necessarily contradictory.

          This might be relevant:

          • 2. Re: reboot cause
            They both came to the same conclusion:

            <pre>"forced crash dump initiated at user request"</pre>

            It so happens that SCAT decided to give a suggestion as to the cause:

            <pre>==== system appears to have been Stop-A'ed ====</pre>

            In this case, it's not correct.
            • 3. Re: reboot cause
              Thank you for your answers!

              Is there a way (for newbies like me) to make SCAT "say" that oprocd.bin is the process that initiated the panic?

              Many thanks!