Skip to contentSkip to conversation

HELP 13 disk raid 5 not starting

DISCUSSION Jim Buttafuoco 3 participants 6 messages Nov 15, 2005 MD RAID

lore
Share & tools
lore
#0 lore raw

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty. When I do the following command to start the raid set (md0) I get an error. Any ideas on how to recover?

This is a debian sarge system running kernel 2.6.8-1-686-smp and mdadm version v1.4.0 - 29 Oct 2003

Thanks
Jim

# mdadm --assemble --scan
mdadm: /dev/md0 assembled from 6 drives and 1 spare - not enough to start the array.

my /etc/mdadm/mdadm.conf looks like

DEVICE partitions
ARRAY /dev/md0 level=raid5 num-devices=12 UUID=c9669a6b:cdc5b9d4:1df0802c:86d53e17

and /proc/mdstat looks like

Personalities : [raid5]
md0 : inactive sdb1[0] sdm1[12] sdn1[11] sdl1[10] sdk1[9] sdj1[8] sdi1[7] sdh1[6] sdg1[5] sdf1[4] sde1[3] sdd1[2] sdc1[1]

      3173710592 blocks

unused devices: <none>

#1 lore raw
Quoted 4 lines · On Tuesday November 15, [email protected] wrote:

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty.

And important question to answer is 'how did this happen'?

Quoted 3 lines

When I do the following command
to start the raid set (md0) I get an error. Any ideas on how to recover?

Add '--force' to the 'mdadm --assemble' command. This tells mdadm to try really hard to assemble the array, modifying info in the super blocks if necessary.
Be aware that though doing this will normally give you a working array, there may be data corruption within the array (it depends somewhat on the answer to that first important question). I would recommend at least an 'fsck' if that is practical.

The array will be assembled degraded. You will need to add in a spare if you are happy that the data is sufficiently intact.

NeilBrown

#2 lore raw

I believe there were data access errors on the console (scrolling to fast to read). I will try the force and see what happends.

---------- Original Message -----------
From: Neil Brown <[email protected]>
To: [email protected]
Cc: [email protected]
Sent: Wed, 16 Nov 2005 10:16:22 +1100
Subject: Re: HELP 13 disk raid 5 not starting

Quoted 20 lines · On Tuesday November 15, [email protected] wrote:
Quoted 4 lines

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty.

And important question to answer is 'how did this happen'?

Quoted 3 lines

When I do the following command
to start the raid set (md0) I get an error. Any ideas on how to recover?

Add '--force' to the 'mdadm --assemble' command. This tells mdadm to try really hard to assemble the array, modifying info in the super blocks if necessary.
Be aware that though doing this will normally give you a working array, there may be data corruption within the array (it depends somewhat on the answer to that first important question). I would recommend at least an 'fsck' if that is practical.

Quoted 8 lines

The array will be assembled degraded. You will need to add in a spare if you are happy that the data is sufficiently intact.

NeilBrown
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

------- End of Original Message -------

#3 lore raw
In reply to Neil Brown
Quoted 20 lines · Neil Brown
Quoted 4 lines · On Tuesday November 15, [email protected] wrote:

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty.

And important question to answer is 'how did this happen'?

Quoted 3 lines

When I do the following command
to start the raid set (md0) I get an error. Any ideas on how to recover?

Add '--force' to the 'mdadm --assemble' command. This tells mdadm to try really hard to assemble the array, modifying info in the super blocks if necessary.
Be aware that though doing this will normally give you a working array, there may be data corruption within the array (it depends somewhat on the answer to that first important question). I would recommend at least an 'fsck' if that is practical.

Quoted 2 lines

The array will be assembled degraded. You will need to add in a spare if you are happy that the data is sufficiently intact.

Before you do anything that might make things worse (--force is a good thing to try, I'm sure, but sometimes the good thing to try causes further problems, EG sometimes fsck goes kerflooey on ya), you might want to see about getting an image backup of all the drives in the RAID array to some other system, if you have the space. If your array wasn't that full, you may be able to compress the image backups to good effect.

#4 lore raw

The disks contain a ~2TB Postgresql database so i was unable to copy to another system. The force worked as suggested by Neil.

Thanks for the reply
Jim

---------- Original Message -----------
From: Dan Stromberg <[email protected]>
To: Neil Brown <[email protected]>
Cc: [email protected], [email protected], [email protected]
Sent: Tue, 15 Nov 2005 15:27:04 -0800
Subject: Re: HELP 13 disk raid 5 not starting

Quoted 31 lines · Neil Brown
Quoted 20 lines · On Tuesday November 15, [email protected] wrote:
Quoted 4 lines

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty.

And important question to answer is 'how did this happen'?

Quoted 3 lines

When I do the following command
to start the raid set (md0) I get an error. Any ideas on how to recover?

Add '--force' to the 'mdadm --assemble' command. This tells mdadm to try really hard to assemble the array, modifying info in the super blocks if necessary.
Be aware that though doing this will normally give you a working array, there may be data corruption within the array (it depends somewhat on the answer to that first important question). I would recommend at least an 'fsck' if that is practical.

Quoted 2 lines

The array will be assembled degraded. You will need to add in a spare if you are happy that the data is sufficiently intact.

Before you do anything that might make things worse (--force is a good thing to try, I'm sure, but sometimes the good thing to try causes further problems, EG sometimes fsck goes kerflooey on ya), you might want to see about getting an image backup of all the drives in the RAID array to some other system, if you have the space. If your array wasn't that full, you may be able to compress the image backups to good effect.

Quoted 4 lines

-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

------- End of Original Message -------

#5 lore raw
In reply to Neil Brown

Neil,

Thanks for the reply, the --force worked great, md0 is syncing now, I will run testing against my database once the sync completes in 400 minutes.

Jim

---------- Original Message -----------
From: Neil Brown <[email protected]>
To: [email protected]
Cc: [email protected]
Sent: Wed, 16 Nov 2005 10:16:22 +1100
Subject: Re: HELP 13 disk raid 5 not starting

Quoted 20 lines · On Tuesday November 15, [email protected] wrote:
Quoted 4 lines

all,

I have a 13 disk raid 5 set with 4 disks marks as "clean" and the rest marked as dirty.

And important question to answer is 'how did this happen'?

Quoted 3 lines

When I do the following command
to start the raid set (md0) I get an error. Any ideas on how to recover?

Add '--force' to the 'mdadm --assemble' command. This tells mdadm to try really hard to assemble the array, modifying info in the super blocks if necessary.
Be aware that though doing this will normally give you a working array, there may be data corruption within the array (it depends somewhat on the answer to that first important question). I would recommend at least an 'fsck' if that is practical.

Quoted 8 lines

The array will be assembled degraded. You will need to add in a spare if you are happy that the data is sufficiently intact.

NeilBrown
-
To unsubscribe from this list: send the line "unsubscribe linux-raid" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

------- End of Original Message -------