Another ST3000DM001 fail (Spin, Recognized, 100% I/O errors)

[Moff Tigriss]

Hi,

Sometimes, you have a bad day, and sometimes you get a premium one.

I have a linux RAID 5 with 6 ST3000. The last night, one drive crash (clicking, SMART scrambled... A good death). No problem, i have a spare disk that I planned for upgrade to RAID 6 (litteraly in two days... damn you, Murphy).
But during resync (~70%), something goes wrong, and the data rate drop to a small 1MB/s, or even less, and rsync fail.
Emmergency state, stop everything, and i try to make a ddrescue clone. It work, but after hours of realy slow copy, the drive refuse to work : I/O error for every single sector.
After a reboot, i was able to restart the ddrescue, and the failed ones pass (good news, not grinded plates \o/), but after à couple of minutes, I/O errors again.
Stop everything, pull off the drive, and... take a deep breath.

So, there is the situation :
- Model : ST3000DM001
- FW : CC26
- PN : 1CH166-302
- Square sticker on the PCB : 1332 H

- Viewable in BIOS, showing correct informations
- SMART alert on boot
- If plugged at boot, i can't pass the POST (freeze, read led cycling).
- If plugged after boot, i can view things with linux tools (SMART, hdparm... but it's really slow).
- After time, 100% I/O errors (even for hdparm).
- Spin normaly (no on/off)
- No clicks or weird noises

Historicaly, the disc was in a previous version of the server. On 4 disks, one fail miserabely (i/o errors, SMART not viewable, and finaly not even shown by the BIOS. It was successfully RMAed. I changed the hardware after that, and everything was ok since), one crash (last night), and the one in question here. Ho, and the system disc goes wrong too on the old server. I think something was really wrong with the power unit.
The last of the four is in ddrescue, for security, with no problem actually (40%). (Something to know, P/N, FW ans PCB are se same. Good to know if a PCB+ROM transplant is on the air).

TL;DR : RAID 5 with two failed disks, one really dead, one seems ok physicaly but make I/O errors after time. I need to recover the raw datas of the last one, to be able to reconstruct the RAID 5.

So, i have some knownledges, but this is far from my level (and i have no tools). The data stored are critical (hopefully, i have a 1 month old 3TB backup of the really really vital ones), of course.
Any suggestions are welcome

Also, i'm in France. I know there is some pretty good DR in the market here, but i have no contact to help me or even tell me where to go.
Money is not a problem, and if i need to make 1000km to work with someone/team, it's not a problem too (and actually, i prefer move the hdd by myself ^^).

Thanks you for your answers

[Dmitri]

I have a linux RAID 5 with 6 ST3000 ... I planned for upgrade to RAID 6 ... But during resync (~70%), something goes wrong

What do you call a "resync"? Is it a rebuild or some RAID 5 -> RAID 6 migration procedure or something else?

[Moff Tigriss]

I've re-read your post it looks like that you are considering to send it to a professional firm already !

Hopefully there will be someone from France reading this forum (DR professional) that are willing to take a look at the drive for a reasonable price !

This is indeed the best decision to make on your case.

Thanks for your advices (very interesting links !) I stopped the process as soon it has shown a degradation of it's allready-bad-condition (not my first dying disk ).
I send some requests to local DR firms (with labs), but now i need to wait monday.

All I hope is that my RAID 5 don't mess with the clone after all... Yay, imbricated-stress !

I have a linux RAID 5 with 6 ST3000 ... I planned for upgrade to RAID 6 ... But during resync (~70%), something goes wrong

What do you call a "resync"? Is it a rebuild or some RAID 5 -> RAID 6 migration procedure or something else?

My bad, I code to much. It's a rebuild.
R5-> R6 was planned, but not launched yet.

[w.simon]

Hello

Where are you located in France ?

[ShaneWard]

How is all your drives heat wise?Iit could be an overheating issue damaging things and hours of cloning will cause a drive to heat up and cause damage to PCB board and internal components. Seagate drives get very hot vary fast.

[Moff Tigriss]

Hello

Where are you located in France ?

Near Paris, but I don't care if I need to move.

How is all your drives heat wise?Iit could be an overheating issue damaging things and hours of cloning will cause a drive to heat up and cause damage to PCB board and internal components. Seagate drives get very hot vary fast.

Never more than 40-42°C during full and long term activity. There is two 120cm fans, at constant RPM, on the side of disks on the actuel case (and the previous hardware was a little more efficient for cooling).

[seed.helper]

Mate this is only one user talking but if you want things to last buy an*ther brand lol... seagate needs to put their act together and stop playing with the costumers

[Moff Tigriss]

they have the firmware bug on 7200.11 (bsy and lba 0)

Ha, i remember when I "fixed" a bricked 750Gb 7200.11. Serial USB, deleting the G-List, the SMART... and, because it was a NTFS disk, the metafile $BADCLUS. Fun times (it worked ).

Actually, what I need to choose to rebuild my hadware, after this mess ? (I look for 3Tb disks only)

I continue search for DR firms. Tomorrow, I will request some quotations, and contact w.simon.

Ho, and after logs reading, i think the first drive (the one who make noise) can be still readable (probably with heads change...). The RAID worked only for somes hours without it, and only 15-20 files were modified during this time (and not even created or size changed). It's a potential try if cloning don't work.

[seed.helper]

i have peace of mind with my several wd disks or even samsung before seagate.

The first try with seagate was in 2009 and the first 1 tb drive and developed bad sectors with a week of use...

And like other user says they are in recovery business now and i dont trust their stuff. They shouldnt make hard drives and be on recovery business at the same time

[Dmitri]

My bad, I code to much. It's a rebuild.
R5-> R6 was planned, but not launched yet.

Then I'd focus on resolving this failed RAID rebuild issue prior to reviving the clicking drive.
I haven't seen the data, but if rebuild wasn't forced and was launched only once, then I believe I should be able to help remotely.
Please check your PM.