Problem with Software RAID5

Problem with Software RAID5

[Lars Schimmer]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi!

I´ve got a problem with a software raid5.
PC runs debian sarge/sid/etch mix with 2.6.16 self built and mdadm mdadm
2.5.3.git200608202239-7.
1 of 6 SATA 400GB drives failes and I rebooted the PC.
After reboot RAID was resyncing.
But again the HD died and the PC rebooted again, I pulled off the bad HD
and now the RAID5 won´t resync again.
I built a 2.6.17 kernel and replaced mdadm to  mdadm 2.5.4-1 and still
RAID5 won´t resync again:
cat /proc/mdstat
Personalities : [linear] [raid0] [raid1] [raid10] [raid5] [raid4]
[raid6] [multipath] [faulty]
md0 : inactive sda1[0] sde1[5] sdd1[4] sdc1[3] sdb1[1]
      1953543680 blocks

unused devices: <none>

 mdadm --detail /dev/md0
/dev/md0:
        Version : 00.90.03
  Creation Time : Fri May 12 16:10:24 2006
     Raid Level : raid5
    Device Size : 390708736 (372.61 GiB 400.09 GB)
   Raid Devices : 6
  Total Devices : 5
Preferred Minor : 0
    Persistence : Superblock is persistent

    Update Time : Tue Oct 17 14:11:56 2006
          State : active, degraded, Not Started
 Active Devices : 5
Working Devices : 5
 Failed Devices : 0
  Spare Devices : 0

         Layout : left-symmetric
     Chunk Size : 256K

           UUID : 5ce125ae:b76d7567:a531953b:fbba92fc
         Events : 0.2818447

    Number   Major   Minor   RaidDevice State
       0       8        1        0      active sync   /dev/sda1
       1       8       17        1      active sync   /dev/sdb1
       2       0        0        2      removed
       3       8       33        3      active sync   /dev/sdc1
       4       8       49        4      active sync   /dev/sdd1
       5       8       65        5      active sync   /dev/sde1

mdadm --stop /dev/md0
mdadm: stopped /dev/md0

sinope:~# mdadm --assemble /dev/md0
mdadm: failed to RUN_ARRAY /dev/md0: Input/output error

Any hints? tips?


MfG,
Lars Schimmer
- --
- -------------------------------------------------------------
TU Graz, Institut für ComputerGraphik & WissensVisualisierung
Tel: +43 316 873-5405       E-Mail: l.schimmer@cgv.tugraz.at
Fax: +43 316 873-5402       PGP-Key-ID: 0x4A9B1723
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iD8DBQFFNfZgmWhuE0qbFyMRAqBBAJoCcE4gMx83NQl8pksSqgEpBHWNiACfTpKr
NVHtinnXRPIbY2Rfv3BUC0s=
=svsf
-----END PGP SIGNATURE-----
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Problem with Software RAID5

[Neil Brown]

On Wednesday October 18, l.schimmer@cgv.tugraz.at wrote:
> 
> Any hints? tips?
> 

If you crash and lose a drive at the same time you could have
undetectable data corruption.  If you are willing to accept that use
'--force'

  mdadm --assemble --force /dev/md0

Hopefully mdadm-2.6 will give a more useful error message here.

NeilBrown

Raid5 or 6 here... ?

[Gordon Henderson]

Heres an oddity - Just built a server with 15 external disks over 2 SAS
channels and I've noticed that the kernel is saying it's RAID5 rather than
RAID6 ...

Hard to explain what I mean in words, but:

  bertha:~# cat /proc/mdstat
  Personalities : [raid0] [raid1] [raid6] [raid5] [raid4]
  md9 : active raid6 sdh1[14] sdo1[13] sdg1[12] sdn1[11] sdf1[10] sdm1[9] sde1[8] sdl1[7] sdd1[6] sdk1[5] sdc1[4] sdj1[3] sdb1[2] sdi1[1] sda1[0]      6341993216 blocks level 6, 128k chunk, algorithm 2 [15/15]
	[UUUUUUUUUUUUUUU]

  md1 : active raid1 sdq1[1] sdp1[0]
      497856 blocks [2/2] [UU]

  md2 : active raid1 sdq2[1] sdp2[0]
      2000000 blocks [2/2] [UU]

  md3 : active raid1 sdq3[1] sdp3[0]
      2000000 blocks [2/2] [UU]

  md5 : active raid1 sdq5[1] sdp5[0]
      73240192 blocks [2/2] [UU]

  md6 : active raid1 sdq6[1] sdp6[0]
      78509504 blocks [2/2] [UU]

  unused devices: <none>

and:

  bertha:~# ps ax | grep raid
   1022 ?        S<     0:00 [md6_raid1]
   1026 ?        S<     0:00 [md5_raid1]
   1030 ?        S<     0:00 [md3_raid1]
   1034 ?        S<     0:00 [md2_raid1]
   1038 ?        S<     0:06 [md1_raid1]
   1747 ?        S<   724:25 [md9_raid5]

Note the raid5 at the bottom, even though /proc/mdstat says it's raid-6.
It was created with:

  mdadm --create /dev/md9 -n15 -l6 --chunk=128    \
        /dev/sda1 /dev/sdi1             \
        /dev/sdb1 /dev/sdj1             \
        /dev/sdc1 /dev/sdk1             \
        /dev/sdd1 /dev/sdl1             \
        /dev/sde1 /dev/sdm1             \
        /dev/sdf1 /dev/sdn1             \
        /dev/sdg1 /dev/sdo1             \
        /dev/sdh1

Is this right? I've been thrashing it all week and just notice this. On
other servers I have using RAID-6, I get what I expect in the 'ps' output.
I was just about to start removing disks from it to test rebuild, etc...
Now not so sure!

It's kernel 2.6.18 and
  mdadm - v1.9.0 - 04 February 2005
(Debian Sarge)

Finally:

  bertha:~# mdadm --detail /dev/md9
  /dev/md9:
        Version : 00.90.03
  Creation Time : Tue Oct 17 14:35:06 2006
     Raid Level : raid6
     Array Size : 6341993216 (6048.20 GiB 6494.20 GB)
    Device Size : 487845632 (465.25 GiB 499.55 GB)
   Raid Devices : 15
  Total Devices : 15
Preferred Minor : 9
    Persistence : Superblock is persistent

    Update Time : Tue Oct 24 06:26:03 2006
          State : clean
 Active Devices : 15
Working Devices : 15
 Failed Devices : 0
  Spare Devices : 0

     Chunk Size : 128K

           UUID : c1a3973d:c46b7356:402fd93c:9778e5a5
         Events : 0.8

    Number   Major   Minor   RaidDevice State
       0       8        1        0      active sync   /dev/sda1
       1       8      129        1      active sync   /dev/sdi1
       2       8       17        2      active sync   /dev/sdb1
       3       8      145        3      active sync   /dev/sdj1
       4       8       33        4      active sync   /dev/sdc1
       5       8      161        5      active sync   /dev/sdk1
       6       8       49        6      active sync   /dev/sdd1
       7       8      177        7      active sync   /dev/sdl1
       8       8       65        8      active sync   /dev/sde1
       9       8      193        9      active sync   /dev/sdm1
      10       8       81       10      active sync   /dev/sdf1
      11       8      209       11      active sync   /dev/sdn1
      12       8       97       12      active sync   /dev/sdg1
      13       8      225       13      active sync   /dev/sdo1
      14       8      113       14      active sync   /dev/sdh1

Gordon

Re: Raid5 or 6 here... ?

[David Greaves]

Gordon Henderson wrote:
>    1747 ?        S<   724:25 [md9_raid5]
> 
> It's kernel 2.6.18 and

Wasn't the module merged to raid456 in 2.6.18?

Are your mdx_raid6's earlier kernels. My raid 6 is on 2.7.17 and says _raid6

Could it be that the combined kernel thread is called mdX_raid5

David

Re: Raid5 or 6 here... ?

[Gordon Henderson]

On Tue, 24 Oct 2006, David Greaves wrote:

> Gordon Henderson wrote:
> >    1747 ?        S<   724:25 [md9_raid5]
> >
> > It's kernel 2.6.18 and
>
> Wasn't the module merged to raid456 in 2.6.18?

Ah, was it? I might have missed that...

> Are your mdx_raid6's earlier kernels. My raid 6 is on 2.7.17 and says _raid6

Yes - earlier ones are < 2.6.18.

> Could it be that the combined kernel thread is called mdX_raid5

If this is the case, then yes! It just seemed a bit odd to me ...

Cheers,

Gordon

Re: Raid5 or 6 here... ?

[David Greaves]

David Greaves wrote:
> Gordon Henderson wrote:
>>    1747 ?        S<   724:25 [md9_raid5]
>>
>> It's kernel 2.6.18 and
> 
> Wasn't the module merged to raid456 in 2.6.18?
> 
> Are your mdx_raid6's earlier kernels. My raid 6 is on 2.7.17 and says _raid6
> 
> Could it be that the combined kernel thread is called mdX_raid5
> 

Yup
raid5.c now handles 45 and 6 and says:
		mddev->thread = md_register_thread(raid5d, mddev, "%s_raid5");

I think I may actually be able to patch that...

David