RAID strangeness

RAID strangeness

[Bruce Pinsky]

I have a four disk RAID 5 array running on my SPARC 10 running Debian 
3.0+ (2.4.20 kernel) and RAID raidtools-0.90 with this raidtab:

raiddev /dev/md0
         raid-level      5
         nr-raid-disks   4
         nr-spare-disks  0
         persistent-superblock 1
         parity-algorithm        left-symmetric
         chunk-size      32
         device          /dev/sdc1
         raid-disk       0
         device          /dev/sdd1
         raid-disk       1
         device          /dev/sde1
         raid-disk       2
         device          /dev/sdf1
         raid-disk       3

It was up and running for more than a month until this weekend.  My 
system rebooted unexpectedly (looks like a power failure at this point) 
and when it rebooted I got this message:

raid5: measuring checksumming speed
    8regs     :    60.000 MB/sec
    32regs    :    62.400 MB/sec
    SPARC     :    81.600 MB/sec
raid5: using function: SPARC (81.600 MB/sec)
md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
md: raid5 personality registered as nr 4
md: could not lock sdc1, zero-size? Marking faulty.
md: could not import sdc1!
md: autostart sdc1 failed!

Investigation has led me to think that my partition tables have 
disappeared based on this:

son-of-spike:/usr/share/doc/raidtools2#  fdisk -l

Disk /dev/sda (Sun disk label): 3 heads, 415 sectors, 14382 cylinders
Units = cylinders of 1245 * 512 bytes

    Device Flag    Start       End    Blocks   Id  System
/dev/sda1             1      7191   4475775   83  Linux native
/dev/sda2          7192     14382   4475775   83  Linux native
/dev/sda3             0     14382   8952795    5  Whole disk

Disk /dev/sdb (Sun disk label): 11 heads, 108 sectors, 3528 cylinders
Units = cylinders of 1188 * 512 bytes

    Device Flag    Start       End    Blocks   Id  System
/dev/sdb1             1       259    153252   83  Linux native
/dev/sdb2  u        260       701    261954   82  Linux swap
/dev/sdb3             0      3528   2095632    5  Whole disk
/dev/sdb7           702      3528   1678644   83  Linux native

Disk /dev/sdc: 64 heads, 32 sectors, 8683 cylinders
Units = cylinders of 2048 * 512 bytes

Disk /dev/sdc doesn't contain a valid partition table

Disk /dev/sdd: 64 heads, 32 sectors, 8683 cylinders
Units = cylinders of 2048 * 512 bytes

Disk /dev/sdd doesn't contain a valid partition table

Disk /dev/sde: 64 heads, 32 sectors, 8683 cylinders
Units = cylinders of 2048 * 512 bytes

Disk /dev/sde doesn't contain a valid partition table

Disk /dev/sdf: 64 heads, 32 sectors, 8683 cylinders
Units = cylinders of 2048 * 512 bytes

Disk /dev/sdf doesn't contain a valid partition table



Thankfully nothing too important on the array yet, but I'm concerned for 
when I do put stuff on it.  Anyone seen a situation where the array 
craps and the partition tables are missing?   Can it be recovered? 
Something wrong with my RAID setup?

Thanks for the help...

-- 
==========
bep

Re: RAID strangeness

[Andre' Breiler]

Hi Bruce,

On Tue, 27 May 2003, Bruce Pinsky wrote:

> I have a four disk RAID 5 array running on my SPARC 10 running Debian
> 3.0+ (2.4.20 kernel) and RAID raidtools-0.90 with this raidtab:

[snip] as it has nothing to do with raid.

> It was up and running for more than a month until this weekend.  My
> system rebooted unexpectedly (looks like a power failure at this point)
> and when it rebooted I got this message:
>
> raid5: measuring checksumming speed
>     8regs     :    60.000 MB/sec
>     32regs    :    62.400 MB/sec
>     SPARC     :    81.600 MB/sec
> raid5: using function: SPARC (81.600 MB/sec)
> md: md driver 0.90.0 MAX_MD_DEVS=256, MD_SB_DISKS=27
> md: raid5 personality registered as nr 4
> md: could not lock sdc1, zero-size? Marking faulty.
> md: could not import sdc1!
> md: autostart sdc1 failed!
>
> Investigation has led me to think that my partition tables have
> disappeared based on this:

> Disk /dev/sdc: 64 heads, 32 sectors, 8683 cylinders
> Units = cylinders of 2048 * 512 bytes
>
> Disk /dev/sdc doesn't contain a valid partition table

I guess your partitions started from the beginning of the disk.

> Thankfully nothing too important on the array yet, but I'm concerned for
> when I do put stuff on it.  Anyone seen a situation where the array
> craps and the partition tables are missing?   Can it be recovered?
> Something wrong with my RAID setup?

Yes I've seen that before but it isn't a problem with raid but sparc/linux
.

If you start your partition from the beginning of your drive then
the partition table will be overwritten when data is written at the
beginning of the first partition.

You can recover your partition table (guess the bounderies) but make sure
you don't write it to the drive or you will overwrite a bit of the first
partition.

Hope that helps,
 Andre'
--