Re: Can not dump on raid dev?

From: Andrew Kinney (andykinney_at_advantagecom.net)
Date: 09/29/03

  • Next message: Tony A, Fields: "nVidia nForce2 potential owners please read (take two)"
    To: "Kang Liu" <liukang@bjpu.edu.cn>
    Date: Mon, 29 Sep 2003 12:14:50 -0700
    
    

    On 29 Sep 2003, at 15:35, Kang Liu wrote:

    > Hi,
    > I'm trying to save a crashdump into /var/crash but it fails.
    > Here is my config files:
    > # dmesg | grep amr
    > amr0: <LSILogic MegaRAID> mem 0xf0000000-0xf7ffffff irq 3 at device 0.0 on pci3
    > amr0: <LSILogic PERC 3/DC> Firmware 1.92, BIOS 3.31, 128MB RAM
    > amrd0: <LSILogic MegaRAID logical drive> on amr0

    <snip>

    > when the system crashes, it doesn't save any crashdump.
    > I've tried the same config on a machine with IDE disk, the crashdump can be saved.
    > Do I miss something, or the crashdump can not be saved on a raid device?
    >

    We had the same trouble with the mlx driver under 4.5, 4.7, and
    4.8. The driver doesn't support crash dumps. I even tried making a
    ccd device since the driver code for ccd appears to support crash
    dumps. Didn't have any luck with that either. Ended up taking an
    old 4GB IDE hard drive and setting it up as a dedicated crash
    dump device.

    Just be sure you don't mount the swap partition you create on it
    because you don't want to compromise the benefits of having a
    RAID if you're using it for redundancy. The only other trick with
    that is that you have to have a BIOS that will allow you to define
    SCSI/RAID as your primary boot device. Some automatically
    assume you want to use IDE for boot if you have and IDE hard drive
    installed. If you're using RAID, you're probably also using a decent
    server board that includes the necessary boot options.

    If memory serves correctly, from what I read in the code, the only
    block device drivers that support crash dumps are IDE and some
    Adaptec SCSI cards. Don't know for sure about those Adaptec
    SCSI cards since we're not using one and you can't always trust
    what you see in the code about dumps as evidenced by the ccd
    driver. I believe it has something to do with it being necessary to
    use BIOS routines to write to the device since by the time you're
    doing a crash dump all your high level drivers are not useable
    anyway.

    Hope that helps.

    Sincerely,
    Andrew Kinney
    President and
    Chief Technology Officer
    Advantagecom Networks, Inc.
    http://www.advantagecom.net

    _______________________________________________
    freebsd-hackers@freebsd.org mailing list
    http://lists.freebsd.org/mailman/listinfo/freebsd-hackers
    To unsubscribe, send any mail to "freebsd-hackers-unsubscribe@freebsd.org"


  • Next message: Tony A, Fields: "nVidia nForce2 potential owners please read (take two)"

    Relevant Pages

    • [git pull] IDE (a.k.a. "zombie") updates #2
      ... Conversion of ide-{cd,floppy} to new IDE debugging macros. ... sgiioc4 host driver fixes. ... drop 'timeout' and 'expiry' arguments from ide_pc_intr ... Add a special packet command request to the tail of the request queue, ...
      (Linux-Kernel)
    • [git patches] IDE updates part #5
      ... Palmchip BK3710 IDE driver ... int stat, err, sense_key; ... + * After each failed packet command we issue a request sense command and retry ...
      (Linux-Kernel)
    • PROBLEM: Buffer I/O error on device hdg1, system freeze.
      ... ide: failed opcode was: unknown ... # APM BIOS Support ... the floppy tape device driver ... PS/2 Mouse Controller at irq 12 ...
      (Linux-Kernel)
    • [git pull] IDE updates [part 2]
      ... Old bast-ide driver removal (we use PATA platform support now) from Ben Dooks. ... Conversion of more host drivers to use struct ide_port_info. ... ide: remove unused XFER_UDMA_SLOW ...
      (Linux-Kernel)
    • [git patches] IDE updates part #3
      ... IDE updates continued... ... move verbose error logging / cdrom.c IOCTLs support ... amd74xx host driver cleanup ... -/* Configuration flags. ...
      (Linux-Kernel)