 EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA
1 edit | ddrescue I'm running DDRescue on a drive and it's been going for 24 hours now it says last successful read was 14 hours ago, it's copied 1.1 TB of data so far. It's a 1.5 TB Drive with about 1.3TB formatted. How long should DDRescue take and will it eventually just finish?
I'm curious as to how long i should expect on average for it to take and at what point do I know it's done? I'm using
ddrescue /dev/sdb /dev/sda drive11.log -r -1 --force
------
Edit: I just got this error Write failed: Broken pipe.
--
Edrick Smith
Independent Film & Broadcast Producer
»edricksmith.com |
|
|
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | Anyone? I did a copy got that error and was told this by Thecus
there is no superblock info on both disks.
mdadm: No md superblock detected on /dev/sda2.
mdadm: No md superblock detected on /dev/sdb2.
--
Edrick Smith
Independent Film & Broadcast Producer
»edricksmith.com |
|
 misiek
join:2000-12-25
Round Lake, IL | reply to Edrick
I wish I could help - perhaps it's a drive issue? In any case, you may want to ask the mods to split your thread between this and Computer Hardware Help sub-forum. User Koitsu seems to know a lot about hard drives. |
|
 BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3
 ·Bell Fibe
| reply to Edrick
Were these disks part of RAID array before?
Linux raid (mdadm) is trying to see if it can find valid raid info, guess the disk partition is set to "fd" linux raid id.
You can mount the partition anyway by forcing the correct fstype on mount or you can change the partition id to "83" and that should fix it.
Why don't you post output of "fdisk -l" before you do any changes? |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | It's part of a thecus array. Do these have shell access?
It's the Thecus NV7000 I want to say. So the partition format is Linux raid |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 Reviews:
·Bell Fibe
| Can you describe what exactly are you trying to accomplish here?
You've obviously cloned a disk that was part of array.
1) Was the disk part of array?
2) If yes, why didn't you just swap the bad disk for new and reconstruct the array using the native array reconstruct tools and data from the remaining disks? (assuming you didn't have failure on more disks in array than the redundancy did allow)
I'm not familiar with Thecus NV7000 but I'd be surprised if it didn't offer failed disk replacement tools. BTW the tools under linux is called mdadm, with it you can remove disk from array, add new one, replace bad one, re-sync the array and more. |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | The array kicked two of the drives due to high relocation. So thecus said to clone the two bad drives bypassing the bad sectors. I tried Actrions True Image but when I get to the destination screen all drive destinations are grayed out |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 | What kind of RAID was it? (RAID1, RAID5, RAID6, ...) How many total disks in it? |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA
2 edits | RAID5 with a spare I believe.
The problem is it was given to me by a customer and the raid setup was already in a critical state. Here's the origional e-mail from Thecus.
-----
1. From the config file, the original RAID system consist of 6 drives in degrade mode, you add disk#2/#3 as spare to rebuild RAID system.
2012/01/18 16:59:32 0.0.0.0 raid_degrade raid5,Tray1,Tray7,Tray4,Tray5,Tray6
2012/01/19 14:39:26 192.168.0.230 conf_raid_add_spare raid5,Tray2,Tray3
2012/01/19 23:30:57 0.0.0.0 raid_healthy raid5,Tray1,Tray7,Tray2,Tray4,Tray5,Tray6,Tray3
2. However, disk#1 and disk#2 having very bad sectors, we suggest disks with reallocated sectors more than 32 be replaced immediately.
Disk#1 Reallocated_Sector_Ct - 49
Disk#2 Reallocated_Sector_Ct - FAILING_NOW 4095
3. Disks continue having I/O errors, therefore RAID system become damaged.
2012/01/20 13:12:05 0.0.0.0 disk_io_fail Tray1
2012/01/20 13:12:05 0.0.0.0 disk_io_fail Tray2
2012/01/20 13:49:54 0.0.0.0 raid_degrade raid5,Tray1,Tray7,Tray4,Tray5,Tray6,Tray3,Tray2
2012/01/20 15:02:03 0.0.0.0 disk_io_fail Tray1
2012/01/20 15:02:03 0.0.0.0 disk_io_fail Tray4
2012/01/20 16:06:12 0.0.0.0 raid_damage raid5,Tray7,Tray5,Tray6,Tray3,Tray2,Tray1,Tray4
3. We suggest you to power down the unit, perform block level backup from the original HDD#1/#2 to another health HDDs using product such as Acronis True Image (»www.acronis.com) or similarly tool (»en.wikipedia.org/wiki/Disk_cloning) and BYPASS the bad sectors.
Once bypass bad sectors backup is completed, insert the cloned disk #1/#2 together with original disk #3~#7, power on system see if RAID system will be up in degrade mode.
-----
I've also attached the config files and log files. If anyone knows how to read them.
When I started the system it was in Degraded mode, however the customer didn't have RAID5+1 configured so it was just striping across all drives. It was degraded but not damaged but it wouldn't allow you to mount the RAID volume because of that. So I was instructed to add a spare via the web interface which then went and rebuilt to the spare drive. The status then went to Healthy and I was now able to start transferring files. A few hours later it went to Damaged.
Is there a way to force the RAID system to mount even with the SMART status giving that error? If I scan the drive I get no bad sectors with the scan.
--
Edrick Smith
Independent Film & Broadcast Producer
»edricksmith.com |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | reply to Edrick
Also from what I can determine I have no shell access to the device unless I install mods for it to enable shell access. |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 Reviews:
·Bell Fibe
| reply to Edrick
This is a tough call. I can't really tell from the data available if this is RAID5 or RAID6. I'm going to relay on the provided info it's RAID5, which means you can loose one drive out of all and still have data in tact. Loosing two drives will cause data loss.
I'd personally would first try to reassemble the array with the existing drives (use --force in mdadm command) and if that succeeds then copy the data over to healthy drive.
If you don't have shell access and can't get one you can always pull the drives to another linux machine (or live linux CD) and do it from there.
--
If the --force --assemble fails then another option would be what's already suggested, dd the drive over to new one.
If you're missing superblock you need again to get on command line.
You need to get UUID from one of the existing drives (mdadm --examine /dev/sdX (replace X with each drive letter)), or from /etc/mdadm.conf (can be in different location) and then try to reassemble like this
mdadm --assemble /dev/md0 --uuid=your_UUID_HERE
If you're lucky you'll see something like this
mdadm: /dev/md0 has been started with 6 drives.
If all that fails try to re-create the array as per steps here »www.storageforum.net/forum/showt···p?t=5890
EDIT: And next time your client should replace failing drive as soon as they get the system alert, not wait until another gets bad (I had to say it eve though it's obious  ) |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | I was under the impression that if a RAID controller for example fails you need to get an exact match of a controller other wise the drives wont build the raid volume. I'm relatively new to RAID its self.
From what I gather, the system originally had 5 Data Drives, 1 Spare and one unused drive.
So:
Drive 1, Drive 4, Drive 5, Drive 6, Drive 7 comprised the data.
Drive 2 was the spare.
Drive 1 and Drive 2 failed SMART due to relocations.
I added Drive 3 into the configuration as another spare. The volume then rebuilt and was mounted in a Healthy status. Then after a few hours it failed and went to damaged.
So Drive 1 is the one that has basically croaked the RAID integrity.
Should I be able to put a healthy drive into Tray 2 or 3 and have it rebuild. What I'm curious about is I understand if there's a drive failure where it's got bad sectors or hardware failure I'd be out of luck. But since it just relocated sectors and 35 is the limit and it's only at 43, shouldn't I be able to add new clean drives to Tray 2 or 3 and have it rebuild?
--
Edrick Smith
Independent Film & Broadcast Producer
»edricksmith.com |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | reply to Edrick
Tried installing the module for SSH but it says I have no RAID setup so it would seem I can't install modules with out the RAID volume. |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 Reviews:
·Bell Fibe
| reply to Edrick
said by Edrick:I was under the impression that if a RAID controller for example fails you need to get an exact match of a controller other wise the drives wont build the raid volume.
In a case of true HW raid this would be true. However what you have there is software raid running on linux. The management tool is mdadm.
said by Edrick:From what I gather, the system originally had 5 Data Drives, 1 Spare and one unused drive.
So:
Drive 1, Drive 4, Drive 5, Drive 6, Drive 7 comprised the data.
Drive 2 was the spare.
Drive 1 and Drive 2 failed SMART due to relocations.
I added Drive 3 into the configuration as another spare. The volume then rebuilt and was mounted in a Healthy status. Then after a few hours it failed and went to damaged.
OK, it was RAID5 with 5 drives originally. Drive 2 failed but it was a spare so had no effect on the issue. Remove it and toss to trash.
What you should have done is to remove Drive 1 and replace with brand new drive and rebuild the array.
How did you add the Drive 3? As replacement for Drive 2 or replacement for Drive 1?
If the former then just trash it as it's replacement for the spare and if it was not added to the array it has no effect. If the latter then I'd re-instert Drive 1 and try to add brand new drive, rebuild to healthy, then trash Drive 1 and replace with new one.
To summ it up. RAID5 is array of N+1 drives, so in your scenario you can survive with 4 out of 5. Spare has no effect unless is added to the array as replacement for failed drive at which point it's not a spare anymore but active raid member. |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | The problem right now is that both Drive 1 (the original) and Drive 2 (the spare) have failed. It should of used Drive 3 as the new drive since that was then the spare. It said it rebuilt once I added Drive 3 but now I can't get anything. I suppose I'll have to keep trying to get Thecus support on the line to explain more. I can't access SSH without a running RAID Mount.
Thanks for the help |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 | If it did rebuild but you don't see anything you may need to do filesystem check. (linux command fsck). This should be available via GUI.
What does it currently say about the RAID health? |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | Says there's no RAID volume, according to thecus the problem is that Drive 1 is too degraded for the RAID system to rebuild onto the spare disk. So disk two only has partial data. They said I need to clone Disk 1 to a new disk verify the partition table and call them and they'll remotely force the system to bring the volume back online. They said I cannot do this. (assuming unless I had shell access somehow then I could). So they must have a way to get in them selves to take care of it.
--
Edrick Smith
Independent Film & Broadcast Producer
»edricksmith.com |
|
BranoI hate VogonsPremium,MVM
join:2002-06-25
Burlington, ON
kudos:3 Reviews:
·Bell Fibe
| OK, they'll most likely log in and do --force assembly. You've cloned the drive, don't worry about the superblock (they should fix it ... if not point them to few posts above) ... call them
...let us know how it went |
|
 koitsuPremium,MVM
join:2002-07-16
Mountain View, CA
kudos:14 | reply to misiek
said by misiek:I wish I could help - perhaps it's a drive issue? In any case, you may want to ask the mods to split your thread between this and Computer Hardware Help sub-forum. User Koitsu seems to know a lot about hard drives.
Yep, except I've been intentionally avoiding this thread. 
--
Making life hard for others since 1977.
I speak for myself and not my employer/affiliates of my employer. |
|
EdrickI aspire to tell the story of a lifetimePremium
join:2004-09-11
Woburn, MA | Why's that  |
|
