Kdump Testing

From: Nagesh Sharyathi (sharyathi_at_in.ibm.com)
Date: 04/21/05

  • Next message: Luke Kenneth Casson Leighton: "noddy question involving /dev/vc/0 and /dev/fb/0 on"
    To: linux-kernel@vger.kernel.org, fastboot@lists.osdl.org
    Date:	Thu, 21 Apr 2005 19:26:11 +0530

    I tested the kdump tool on x235 and x206 machines and found this problem
    where on kernel Panic, system instead of booting into the panic kernel
    jumps into BIOS and machine restarts.
    (I have given the hardware specifications at the bottom of the mail)

    - 2.6.12-rc2-mm1
    - kexec-tools-1.101
    - Five kdump user space patches

    Test Procedure:
    - Built first kernel for 1M location with CONFIG_KEXEC enabled.
    - Booted into first kernel with command line options crashkernel=48M@16M.
    - Built second kernel for 16M location with CONFIG_CRASH_DUMP, and
      CONFIG_PROC_VMCORE enabled.
    - Loaded second kernel with following kexec command.

      kexec -p vmlinux-16M --args-linux --crash-dump --append="root=<root-dev>
      init 1"

    - Inserted a module or echo into sysrq-trigger to invoke panic.
    - System jumps into BIOS directly instead of booting into secondary

    Summary Observation:

    - Earlier I was able to make kdump work on x330 machine by removing
    maxcpus=1 (as specified in kdump.txt) option during loading panic kernel,
    through kexec tool. But this work around doesn't seems to work with the
    hardware x235 and x206. On kernel panic machine jumps to BIOS rather than
    to panic kernel without displaying any error message.


    A) Hardware x330:
    - SMP, 2way, Pentium III (Coppermine) 1 GHz, 1.3G RAM
    - Network Interface (e100)
    - Disk I/O
      SCSI storage controller: Adaptec Ultra160
    B)Hardware x235
    - SMP, 2way, Xeon TM 2.8GHz, 1.5g RAM
    - Network Interface (Tigon3)
    - Disk I/O
      SCSI storage controller: IBM Serve RAID
    C)Hardware x206
    - SMP, 1way, Pentium IV 2.8GHz, 2g RAM
    - Network Interface (e1000)
    - Disk I/O
      SCSI storage controller: Adaptec Ultra320

    To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
    the body of a message to majordomo@vger.kernel.org
    More majordomo info at http://vger.kernel.org/majordomo-info.html
    Please read the FAQ at http://www.tux.org/lkml/

  • Next message: Luke Kenneth Casson Leighton: "noddy question involving /dev/vc/0 and /dev/fb/0 on"

    Relevant Pages

    • Re: [PATCH] x86-64: disable the GART before allocate aperture
      ... And having the kexec on panic kernel use the reserved portion? ... How about simply reserving all of it for the base kernel and using soft ... Won't I require IOMMU while I am ...
    • Re: [patch] add kdump_after_notifier
      ... I want to use both kdb and kdump, ... The kexec on panic assumption is that the kernel is broken and we better not ... My stance is that _all_ the RAS tools (kdb, kgdb, nlkd, netdump, lkcd, ...
    • Re: oops in :snd_pcm_oss:resample_expand+0x19c/0x1f0
      ... Thanks for trying out kdump. ... For example, the small script you wrote for saving the dump, is already ... "service kdump start" and initrd will be generated and kdump kernel will ... Currently we can use gdb but only for linearly mapped region. ...
    • RE: Query about kdump_msg hook into crash_kexec()
      ... I explain the usage of kmsg_dumpin enterprise area. ... useful information for investigating kernel crash in case kdump ... kernel doesn't boot. ...
    • Re: Query about kdump_msg hook into crash_kexec()
      ... The whole design of kdump is built on the ... transition to second kernel robust. ... Now think give the control to two fools. ... dump, what are the other purposes of kdump. ...