Re: mdadm raid1

From: Harland Christofferson (debian-user_at_zerocrossings.com)
Date: 11/30/04

  • Next message: Nate Bargmann: "Re: "Grub Hard Disk Error", after installing from sarge floppies" 
    Date: Mon, 29 Nov 2004 21:18:40 -0500
To: debian-user@lists.debian.org

    I think I figured it out but am still looking for suggestions:

    mdadm --stop /dev/md0
    mdadm -C /dev/md0 -l1 -n2 /dev/hda /dev/hdc

    and it looks like the kern.log reflects that the array is being resynchronized.

    Is this correct?

    At Monday, 29 November 2004, Harland Christofferson <debian-user@zerocrossings.
    com> wrote:

    >I had a problem w/ my system crashing and I noticed there was an
    >I/O error to /dev/hda. In a fit of panic, I swapped the IDE cable
    >of /dev/hda and /dev/hdc and reboot. Come to find out, my CPU fan
    >was dying causing the CPU not to work.
    >
    >Anyway ...
    >
    >Upon putting the drives back where they belong, I notice that /dev/hda
    >is kicked out of the raid1 array as faulty. Reading the kern.log,
    >it looks like this is occuring b/c of a superblock update time inconsistency.

    >How can I correct this so that both /dev/hda and /dev/hdc are in
    >the array /dev/md0?
    >
    >Nov 29 20:25:37 trader kernel: md: md driver 0.90.0 MAX_MD_DEVS=256,
    >MD_SB_DISKS=27
    >Nov 29 20:25:37 trader kernel: md: raid1 personality registered as nr 3
    >Nov 29 20:25:37 trader kernel: [events: 00000034]
    >Nov 29 20:25:37 trader kernel: [events: 00000039]
    >Nov 29 20:25:37 trader kernel: md: autorun ...
    >Nov 29 20:25:37 trader kernel: md: considering hdc ...
    >Nov 29 20:25:37 trader kernel: md: adding hdc ...
    >Nov 29 20:25:37 trader kernel: md: adding hda ...
    >Nov 29 20:25:37 trader kernel: md: created md0
    >Nov 29 20:25:37 trader kernel: md: bind<hda,1>
    >Nov 29 20:25:37 trader kernel: md: bind<hdc,2>
    >Nov 29 20:25:37 trader kernel: md: running: <hdc><hda>
    >Nov 29 20:25:37 trader kernel: md: hdc's event counter: 00000039
    >Nov 29 20:25:37 trader kernel: md: hda's event counter: 00000034
    >Nov 29 20:25:37 trader kernel: md: superblock update time inconsistency

    >-- using the most recent one
    >Nov 29 20:25:37 trader kernel: md: freshest: hdc
    >Nov 29 20:25:37 trader kernel: md: kicking non-fresh hda from array!
    >Nov 29 20:25:37 trader kernel: md: unbind<hda,1>
    >Nov 29 20:25:37 trader kernel: md: export_rdev(hda)
    >Nov 29 20:25:37 trader kernel: md: md0: raid array is not clean -
    >- starting background reconstruction
    >Nov 29 20:25:37 trader kernel: md: RAID level 1 does not need chunksize!

    >Continuing anyway.
    >Nov 29 20:25:37 trader kernel: md0: max total readahead window set
    >to 124k
    >Nov 29 20:25:37 trader kernel: md0: 1 data-disks, max readahead per
    >data-disk: 124k
    >Nov 29 20:25:37 trader kernel: raid1: device hdc operational as
    mirror 1
    >Nov 29 20:25:37 trader kernel: raid1: md0, not all disks are operational

    >-- trying to recover array
    >Nov 29 20:25:37 trader kernel: raid1: raid set md0 active with 1
    >out of 2 mirrors
    >Nov 29 20:25:37 trader kernel: md: updating md0 RAID superblock on
    >device
    >Nov 29 20:25:37 trader kernel: md: hdc [events: 0000003a]<6>(write)
    >hdc's sb offset: 40146560
    >Nov 29 20:25:37 trader kernel: md: recovery thread got woken up ...
    >Nov 29 20:25:37 trader kernel: md0: no spare disk to reconstruct
    >array! -- continuing in degraded mode
    >Nov 29 20:25:37 trader kernel: md: recovery thread finished ...
    >Nov 29 20:25:37 trader kernel: md: ... autorun DONE.
    >
    >--
    >To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org
    >with a subject of "unsubscribe". Trouble? Contact listmaster@lists.
    debian.org
    > 

    -- 
To UNSUBSCRIBE, email to debian-user-REQUEST@lists.debian.org 
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
  • Next message: Nate Bargmann: "Re: "Grub Hard Disk Error", after installing from sarge floppies"

    Relevant Pages

    • Re: To serialize or not to serialize?
      ... array, sending smaller portions of data in several times... ... So, when the client asks for some data, the web service queries the ...
      (microsoft.public.dotnet.languages.csharp)
    • Re: SQL 2005 clustered virtual server
      ... We are implementing a 2 node SQL 2005 cluster with SAN array. ... 1 mirror for the logs and 5x mirrors for data in which I am ... further 2x 72GB drives. ... ---rootdisk2 (another drive letter, physical disk resource) ...
      (microsoft.public.windows.server.clustering)
    • Re: RAID options
      ... scenario you can yank the two partners and use them as fallback should ... This would also be my motivation for the additional drives, ... have OS and programs on the mirror and DATA on the RAID5, ... >> Though SBS is IO intensive performance of the array should not be your ...
      (microsoft.public.windows.server.sbs)
    • Re: remove/replace a harddisk in a unsynced raid1?
      ... > disks are running as mirror and the third one is running as spare disk. ... Your array was probably built with 2 RAID devices so anything else will ...
      (comp.os.linux.misc)
    • Re: Drive Replacement with Disk Mirroring
      ... All the mirroring is done in SBS. ... SBS re-equalize the mirror on to the new drive. ... > I understand that you may want to perform a disk replication. ... > controller, create the hardware array, and then format the array. ...
      (microsoft.public.windows.server.sbs)