Assembling RAID 5 array with missing superblock

I just keep having bad luck with my RAID 5 array. One drive (out of 4
WD15EARS drives) was dropped from the array due to powerfailure and
rebuild failed because of bad sectors on another drive. I have been
spending the last few days with ddrescue, copying the data from the
drive with bad sectors to the dropped one. The problem is now that the
drive has bad sectors have spread where the md superblock used to be
on the faulty drive (the very last sectors of the drive, right?), so
how do i force it to assemble without it and with one drive missing?

What drives are by the way the best consumer grade drives to use for
RAID operation?

I really, really hope this is the last time i need to ask you guys for help.

--
Lasse Jensen (fafler at gmail dot com)
--
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: Assembling RAID 5 array with missing superblock

This is a cryptographically signed message in MIME format.

--------------ms020300050408080107020807
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Content-Transfer-Encoding: quoted-printable

On 02/23/11 15:05, Lasse Jensen wrote:
> I really, really hope this is the last time i need to ask you guys for =
help.

Well, just a little bit off-topic, but...:
For me, and probably others as well, it does not harm in any way to keep =
asking=20
questions. I find it interesting and insightful reading some of the answe=
rs to=20
these questions.
So please do ask.

Regards,

Jaap Crezee





--------------ms020300050408080107020807
Content-Type: application/pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature

MIAGCSqGSIb3DQEHAqCAMIACAQExCzAJBgUrDgMCGgUAMIAGCSqGSIb3DQEH AQAAoIIK2jCC
BWkwggNRoAMCAQICAwnAGTANBgkqhkiG9w0BAQUFADB5MRAwDgYDVQQKEwdS b290IENBMR4w
HAYDVQQLExVodHRwOi8vd3d3LmNhY2VydC5vcmcxIjAgBgNVBAMTGUNBIENl cnQgU2lnbmlu
ZyBBdXRob3JpdHkxITAfBgkqhkiG9w0BCQEWEnN1cHBvcnRAY2FjZXJ0Lm9y ZzAeFw0xMTAx
MjgwODEzMDZaFw0xMzAxMjcwODEzMDZaMGIxHjAcBgNVBAMTFUphY29iIEJl bmphbWluIENy
ZXplZTEaMBgGCSqGSIb3DQEJARYLamFhcEBqY3oubmwxJDAiBgkqhkiG9w0B CQEWFWphYXAu
Y3JlemVlQGltdGVjaC5ubDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoC ggEBANnfQCy6
Zrxwlai9XI3qiMDrGK1fFs1OrFjujjwfGCddd/WoXA+5F0dLUtfSxO1YrqTr xRL3WUI/q4Zi
B6AqoPDsp0BcytjTkfILUxwzV2oBkJ29Ksz76fuaroK1IaeOzRujNSzoqKUJ NiykISxyIuQM
odYWpuZxMifCKoRUNtMYzOJH1N8dd4IpY1DK7pfTqm9Z/ToY7J7zNlgsC9DY Kat1iNGO1HN2
hhjPRAqT9Obc/1Bj9BtbObVi7aa88kzbnzNyenkDH+JmfY75I06SHZXOp37E rQcV3SUBeHWS
uftf/xyKekrhDWL4JM3ELaj9r0+TdhbyV4S1bG03D5IU0zMCAwEAAaOCAQ8w ggELMAwGA1Ud
EwEB/wQCMAAwVgYJYIZIAYb4QgENBEkWR1RvIGdldCB5b3VyIG93biBjZXJ0 aWZpY2F0ZSBm
b3IgRlJFRSBoZWFkIG92ZXIgdG8gaHR0cDovL3d3dy5DQWNlcnQub3JnMEAG A1UdJQQ5MDcG
CCsGAQUFBwMEBggrBgEFBQcDAgYKKwYBBAGCNwoDBAYKKwYBBAGCNwoDAwYJ YIZIAYb4QgQB
MDIGCCsGAQUFBwEBBCYwJDAiBggrBgEFBQcwAYYWaHR0cDovL29jc3AuY2Fj ZXJ0Lm9yZzAt
BgNVHREEJjAkgQtqYWFwQGpjei5ubIEVamFhcC5jcmV6ZWVAaW10ZWNoLm5s MA0GCSqGSIb3
DQEBBQUAA4ICAQB8quzWe1GvQwx1HFBCwMbB+0hthvbmDJP3eJRp2ADumRon hrmuRWaH4pqr
wLmT6V65n56WPmO3Vb1FaZuzSP3sozBYJttwZCtptVny1SCI2E1zTpwii+I0 +dmFMD3p5WE6
s+p0pzqdi6uFgKAPtsMh5ZtR/eiSXpzT7MJ892c/t6WvEdNIZHhUan7QhyJf aebrfvdcgdVp
26TYc0AkcLYUrQHNGG9qGykB78VJsPxF25qli2N+p0lQXJNcO4bUE5o7TxF3 Q1BwGrVR+Z0C
CLg1HVXUwUNQc/nClhNjy3MFyy0FsvEBERgAtIpm6XZ2qE7IyiwUcXwQlmP1 BrOqgXPpoxhn
KuYwNlbn08vxPsNo6/HZx3S8gMQNmcDQ7d/r4Zl8FmDxULHu3I+w83tUq8Wf sfbJvry69MMR
Y2OXtd84vai4Bv2Af6cXCBLZtz2aU57EFiIai8dgfa3fbQm5PJPt5u6DYmYR 80GVV4P+oN8Q
/EKOGH27VQQmgJq/CYrwFW1Dac5r0/kZP6z8lKtSRWdq6O6RRK5WpiRieOKn JBm0p0/cB5Om
ytuZkW/wkSRXokbGedKhHRO4Ws/Lr0n9Z7yByVKVkWMGvQiFqv34cfTdduAv 0xi55v/zL91u
HsS1rSlnVWxLHxVeAOgk8KKaeYiIAquyk4n5xOWIBUdEGNUJ8zCCBWkwggNR oAMCAQICAwnA
GTANBgkqhkiG9w0BAQUFADB5MRAwDgYDVQQKEwdSb290IENBMR4wHAYDVQQL ExVodHRwOi8v
d3d3LmNhY2VydC5vcmcxIjAgBgNVBAMTGUNBIENlcnQgU2lnbmluZyBBdXRo b3JpdHkxITAf
BgkqhkiG9w0BCQEWEnN1cHBvcnRAY2FjZXJ0Lm9yZzAeFw0xMTAxMjgwODEz MDZaFw0xMzAx
MjcwODEzMDZaMGIxHjAcBgNVBAMTFUphY29iIEJlbmphbWluIENyZXplZTEa MBgGCSqGSIb3
DQEJARYLamFhcEBqY3oubmwxJDAiBgkqhkiG9w0BCQEWFWphYXAuY3JlemVl QGltdGVjaC5u
bDCCASIwDQYJKoZIhvcNAQEBBQADggEPADCCAQoCggEBANnfQCy6Zrxwlai9 XI3qiMDrGK1f
Fs1OrFjujjwfGCddd/WoXA+5F0dLUtfSxO1YrqTrxRL3WUI/q4ZiB6AqoPDs p0BcytjTkfIL
UxwzV2oBkJ29Ksz76fuaroK1IaeOzRujNSzoqKUJNiykISxyIuQModYWpuZx MifCKoRUNtMY
zOJH1N8dd4IpY1DK7pfTqm9Z/ToY7J7zNlgsC9DYKat1iNGO1HN2hhjPRAqT 9Obc/1Bj9Btb
ObVi7aa88kzbnzNyenkDH+JmfY75I06SHZXOp37ErQcV3SUBeHWSuftf/xyK ekrhDWL4JM3E
Laj9r0+TdhbyV4S1bG03D5IU0zMCAwEAAaOCAQ8wggELMAwGA1UdEwEB/wQC MAAwVgYJYIZI
AYb4QgENBEkWR1RvIGdldCB5b3VyIG93biBjZXJ0aWZpY2F0ZSBmb3IgRlJF RSBoZWFkIG92
ZXIgdG8gaHR0cDovL3d3dy5DQWNlcnQub3JnMEAGA1UdJQQ5MDcGCCsGAQUF BwMEBggrBgEF
BQcDAgYKKwYBBAGCNwoDBAYKKwYBBAGCNwoDAwYJYIZIAYb4QgQBMDIGCCsG AQUFBwEBBCYw
JDAiBggrBgEFBQcwAYYWaHR0cDovL29jc3AuY2FjZXJ0Lm9yZzAtBgNVHREE JjAkgQtqYWFw
QGpjei5ubIEVamFhcC5jcmV6ZWVAaW10ZWNoLm5sMA0GCSqGSIb3DQEBBQUA A4ICAQB8quzW
e1GvQwx1HFBCwMbB+0hthvbmDJP3eJRp2ADumRonhrmuRWaH4pqrwLmT6V65 n56WPmO3Vb1F
aZuzSP3sozBYJttwZCtptVny1SCI2E1zTpwii+I0+dmFMD3p5WE6s+p0pzqd i6uFgKAPtsMh
5ZtR/eiSXpzT7MJ892c/t6WvEdNIZHhUan7QhyJfaebrfvdcgdVp26TYc0Ak cLYUrQHNGG9q
GykB78VJsPxF25qli2N+p0lQXJNcO4bUE5o7TxF3Q1BwGrVR+Z0CCLg1HVXU wUNQc/nClhNj
y3MFyy0FsvEBERgAtIpm6XZ2qE7IyiwUcXwQlmP1BrOqgXPpoxhnKuYwNlbn 08vxPsNo6/HZ
x3S8gMQNmcDQ7d/r4Zl8FmDxULHu3I+w83tUq8WfsfbJvry69MMRY2OXtd84 vai4Bv2Af6cX
CBLZtz2aU57EFiIai8dgfa3fbQm5PJPt5u6DYmYR80GVV4P+oN8Q/EKOGH27 VQQmgJq/CYrw
FW1Dac5r0/kZP6z8lKtSRWdq6O6RRK5WpiRieOKnJBm0p0/cB5OmytuZkW/w kSRXokbGedKh
HRO4Ws/Lr0n9Z7yByVKVkWMGvQiFqv34cfTdduAv0xi55v/zL91uHsS1rSln VWxLHxVeAOgk
8KKaeYiIAquyk4n5xOWIBUdEGNUJ8zGCA5QwggOQAgEBMIGAMHkxEDAOBgNV BAoTB1Jvb3Qg
Q0ExHjAcBgNVBAsTFWh0dHA6Ly93d3cuY2FjZXJ0Lm9yZzEiMCAGA1UEAxMZ Q0EgQ2VydCBT
aWduaW5nIEF1dGhvcml0eTEhMB8GCSqGSIb3DQEJARYSc3VwcG9ydEBjYWNl cnQub3JnAgMJ
wBkwCQYFKw4DAhoFAKCCAegwGAYJKoZIhvcNAQkDMQsGCSqGSIb3DQEHATAc BgkqhkiG9w0B
CQUxDxcNMTEwMjIzMjA1NTU2WjAjBgkqhkiG9w0BCQQxFgQU9LEyqTl4C3NF M6/4/7e2jY2b
usUwXwYJKoZIhvcNAQkPMVIwUDALBglghkgBZQMEAQIwCgYIKoZIhvcNAwcw DgYIKoZIhvcN
AwICAgCAMA0GCCqGSIb3DQMCAgFAMAcGBSsOAwIHMA0GCCqGSIb3DQMCAgEo MIGRBgkrBgEE
AYI3EAQxgYMwgYAweTEQMA4GA1UEChMHUm9vdCBDQTEeMBwGA1UECxMVaHR0 cDovL3d3dy5j
YWNlcnQub3JnMSIwIAYDVQQDExlDQSBDZXJ0IFNpZ25pbmcgQXV0aG9yaXR5 MSEwHwYJKoZI
hvcNAQkBFhJzdXBwb3J0QGNhY2VydC5vcmcCAwnAGTCBkwYLKoZIhvcNAQkQ AgsxgYOggYAw
eTEQMA4GA1UEChMHUm9vdCBDQTEeMBwGA1UECxMVaHR0cDovL3d3dy5jYWNl cnQub3JnMSIw
IAYDVQQDExlDQSBDZXJ0IFNpZ25pbmcgQXV0aG9yaXR5MSEwHwYJKoZIhvcN AQkBFhJzdXBw
b3J0QGNhY2VydC5vcmcCAwnAGTANBgkqhkiG9w0BAQEFAASCAQAvk7LTl4Sx FfogVbBByVPR
WeFUudKh1lCpAE8Ufnm81l8tkXBEo9bQKqnDh7HJDN6DMah83dMzQdFgcuMt YpP3QX3tVuHC
efKhA9wF2gNqZNYnJ6C10h6RkhUllzTApcpfgXiMK3n6MBbbrfsR/lQzKQnV jRbTea0XlGwk
7X0CN0j/01V4bYxwIfLKAOcL+N0Ha3DBnBsIwnxxn9+vqKTlEHfj1z5d+Lvt 1Ou62MRD0CvU
Q4iOHtqHX5/ILosANYTSUJXhlunPuhMMibM7i3xQ2IldU+KifceXc8pcYt5g 4judKqKxXuKt
BrBp6KhENWbkkFR2G4XDyqrVIbFwaymXAAAAAAAA
--------------ms020300050408080107020807--
--
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: Assembling RAID 5 array with missing superblock

On Thu, Feb 24, 2011 at 5:45 AM, wrote:
> On Wed, Feb 23, 2011 at 9:05 PM, Lasse Jensen wrote:
>> What drives are by the way the best consumer grade drives to use for
>> RAID operation?
>
> I've had good luck with Hitachi (used to be IBM) and Samsung lately,
> more so than WD. Hitachi's 5-year warranty (everyone else here being
> only 3) is also an indicator. . .

I used IBM back when they made the infamous DeathStar drives. Havent
brought a IBM/Hitachi drive since. I guess i'm going to feel the same
way about WD drives now. Maybe i should just mix'n'match manufactures.
They can't all be that bad,

> I know Google uses consumer-grade drives (with massive redundancy and
> automated hyper-vigilant sysadmin processes) - anyone know whose?
>
>
>> I just keep having bad luck with my RAID 5 array. One drive (out of 4
>> WD15EARS drives) was dropped from the array due to powerfailure and
>> rebuild failed because of bad sectors on another drive. I have been
>
> Sorry I can't help more but with any such large drives I'd suggest
> sticking with RAID6, plus hot spares if you think you might not notice
> and/or be able to quickly replace a failing member.

That was my plan, but without the hotspare. I have 2 drives i need to
get replaced, and after i hopefully regain access to my data, i'm
going for a 5 drive RAID 6 array.

--
Lasse Jensen (fafler at gmail dot com)
--
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: Assembling RAID 5 array with missing superblock

Hey,

> I used IBM back when they made the infamous DeathStar drives. Havent
> brought a IBM/Hitachi drive since. I guess i'm going to feel the same
> way about WD drives now. Maybe i should just mix'n'match manufactures.
> They can't all be that bad,

Mixing and matching drive manufacturers seems like an excellent idea,
since it makes it less likely that drives will die in droves at the same
time, which makes for a generally saver RAID. On this mailing list we
know that drives fail and expect it, that's why we are on it :)

Cheers,
Iordan
--
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: Assembling RAID 5 array with missing superblock

On 24/02/2011 16:10, Iordan Iordanov wrote:
> Hey,
>
>> I used IBM back when they made the infamous DeathStar drives. Havent
>> brought a IBM/Hitachi drive since. I guess i'm going to feel the same
>> way about WD drives now. Maybe i should just mix'n'match manufactures.
>> They can't all be that bad,
>
> Mixing and matching drive manufacturers seems like an excellent idea,
> since it makes it less likely that drives will die in droves at the same
> time, which makes for a generally saver RAID. On this mailing list we
> know that drives fail and expect it, that's why we are on it :)

Indeed. I started my home server off with a 3-drive RAID-5, with 3
identical drives with consecutive serial numbers. One started giving
building pending sectors within a month, another after 9 months. By now
I have 6 drives in RAID-6 from 3 manufacturers with varying
power-on-hours and no two from the same batch.

Cheers,

John.

--
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: Assembling RAID 5 array with missing superblock

So should I just bite it and build a new array? All the really
important stuff is backed up elsewhere, the rest is, well, just crap i
got of the Internet, you know...

One last thing, I've been thinking about. What copying the superblock
from one of the working drives to the defect one and using a hew
editor to change the drive number and alignment? Is there much device
specific data in the superblock?

On Wed, Feb 23, 2011 at 3:05 PM, Lasse Jensen wrote:
> I just keep having bad luck with my RAID 5 array. One drive (out of 4
> WD15EARS drives) was dropped from the array due to powerfailure and
> rebuild failed because of bad sectors on another drive. I have been
> spending the last few days with ddrescue, copying the data from the
> drive with bad sectors to the dropped one. The problem is now that the
> drive has bad sectors have spread where the md superblock used to be
> on the faulty drive (the very last sectors of the drive, right?), so
> how do i force it to assemble without it and with one drive missing?
>
> What drives are by the way the best consumer grade drives to use for
> RAID operation?
>
> I really, really hope this is the last time i need to ask you guys for help.
>
> --
> Lasse Jensen (fafler at gmail dot com)
>



--
Lasse Jensen (fafler at gmail dot com)
--
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