recover device files from failed RAID5

Status
Not open for further replies.
A

anneranch

Active Member
Joined
Mar 16, 2019
Messages
223
Reaction score
43
Credits
2,078
I have totally lost access to my software RAID5.
I can see the RAID 5 devices , HDD partitions, but file manager won't let me access them.

Is there some other way to recover the contents of the RAID5 device ?

I am currently rebuilding a new RAID5 and I have added one of the failed devices to it,
but the rebuild is taking time... So I will know in abut an hour if the old failed device contents was
copied / added to the new array,

Addendum
Old devices cannot be added to new RAID - they are "busy".
I am looking for a way to unbusy them.

Any help - in that direction - would be appreciated .
 
Last edited:


We would be inclined to give more credit, if you had a backup of some sort.;
 
I really do not need a credit , just more serious attempt to answer the original qustion.
( Lack of knowledge about RAID and squing the disussion to "backup"
is not helping )

Is there some other way to recover the contents of the RAID5 device ?

Hint:
The inaccessible RAID devices are "busy" therefore cannot be accessed.
Perhaps some way to "unbusy " them would help?
 
anneranch said:
I really do not need a credit
Click to expand...
You really do not 'get it' do you?

If in fact you had bothered yourself to have made a backup or similar, it would show a willingness on your part to help yourself.

Instead, the 'message' that comes across is that the good people here are in some way indebted/responsible to provide you with answers. ...Even when your tone is less than friendly, ....instead resorting to demands.

There are certainly simpler, more amicable ways to get help here at Linux.org
 
It would have to be a pretty catrostrophic failure to lose stuff on RAID 5.
How many drives are in your RAID? Even if one goes bad you should still be able to function.
If you replace the bad drive, the parity bits should rebuild your data automatically.

You shouldn't have create an entirely new RAID, just remove the bad drive from the existing RAID.
If you create a new RAID, it could take days to rebuild all the parity. Depending on how much dat ayou have.

Also, use hardware RAID if possible.
 
The problem is NOT hardware failure - this is entirely software RAID5.
The problem is I cannot access any of the devices as files.
I am using lsblk and the devices / partitions only shows as /dev ,
there is no md mount points etc. ( I could post the lsblk , but at this point
of discussion it would be waste of space)
I have tryied to add old partitons to new RAID and getting "device busy".
That tells me the old partitions / devices are still intact,
just the access is restricted.

This all happend when the array was resynching and I lost power.
 
Condobloke said:
You really do not 'get it' do you?

If in fact you had bothered yourself to have made a backup or similar, it would show a willingness on your part to help yourself.

Instead, the 'message' that comes across is that the good people here are in some way indebted/responsible to provide you with answers. ...Even when your tone is less than friendly, ....instead resorting to demands.

There are certainly simpler, more amicable ways to get help here at Linux.org
Click to expand...
Comments like these really do not help to resolve the problem.
 
I am having hard time "attach file" , so here is the actual state of one if the devices/ partition.
The device is active, what is missing in mount point.
The question is :
How do I mount it if all I have as "name" qy-desktop:30
The usual "mdx" is replaced with "qy-desktop:30" and
I have created test mount point , but cannot mount the sdf27

qy@qy-desktop:~$ sudo mkdir /test
[sudo] password for qy:
qy@qy-desktop:~$ sudo mount /dev/sdf27 /test
mount: /test: unknown filesystem type 'linux_raid_member'.

I do not believe changing the file system would be wise.






/dev/sdf27:
Magic : a92b4efc
Version : 1.2
Feature Map : 0x44
Array UUID : d368acc2:824b3dfa:046c4484:76a1ae7f
Name : qy-desktop:30
Creation Time : Sat Dec 25 11:32:28 2021
Raid Level : raid5
Raid Devices : 7

Avail Dev Size : 204786688 sectors (97.65 GiB 104.85 GB)
Array Size : 61400064 KiB (58.56 GiB 62.87 GB)
Used Dev Size : 20466688 sectors (9.76 GiB 10.48 GB)
Data Offset : 13312 sectors
New Offset : 7168 sectors
Super Offset : 8 sectors
State : clean
Device UUID : 1705e33d:dbbeddfd:6ae8fd38:be199743

Reshape pos'n : 18736128 (17.87 GiB 19.19 GB)
Delta Devices : 1 (6->7)

Update Time : Sat Jan 1 13:47:31 2022
Bad Block Log : 512 entries available at offset 136 sectors
Checksum : 1825f00 - correct
Events : 8099

Layout : left-symmetric
Chunk Size : 512K

Device Role : Active device 5
Array State : AAAAAAA ('A' == active, '.' == missing, 'R' == replacing)
 
Here is some more

Please note that I cannot redirect the mdadm scan to the mdadm.conf file

qy@qy-desktop:~$ sudo mdadm --examine --scan >> /etc/mdadm/mdadm.conf
bash: /etc/mdadm/mdadm.conf: Permission denied
qy@qy-desktop:~$ sudo mdadm --examine --scan >> /etc/mdadm/mdadm.conf
bash: /etc/mdadm/mdadm.conf: Permission denied

I cannot access / modify the file manually - it is read only with acces only by root.


Also the array I am trying to recover is marked as ARRAY


qy@qy-desktop:~$ sudo mdadm --examine --scan
[sudo] password for qy:
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
ARRAY /dev/md/1 metadata=1.2 UUID=2fcd6c3c:7536faa9:8eb26bcd:31243bf7 name=z-desktop:1
ARRAY /dev/md/0 metadata=1.2 UUID=f82fd801:3c936a89:9ef1a37f:f6cb9c9d name=qt-desktop:0
ARRAY /dev/md/3 metadata=1.2 UUID=f6e0ad4d:9a46cdd6:2a6c2a1b:a6f6a1e5 name=a-desktop:3
ARRAY /dev/md/10 metadata=1.2 UUID=3fee7915:97923e1c:5165ee2e:3e5191c5 name=a-desktop:10
ARRAY /dev/md/CAT_RAID metadata=1.2 UUID=2f05ad5b:70f5a078:62a355b9:417fdaab name=w-desktop:CAT_RAID
ARRAY /dev/md/0 metadata=1.2 UUID=61f39a05:1987c353:652b86bf:4789cc1c name=w-desktop:0
ARRAY /dev/md/20 metadata=1.2 UUID=cbdba8c2:4b5e8736:66d15b79:71b7d89f name=qe-desktop:20
ARRAY /dev/md/15 metadata=1.2 UUID=c27a45b2:62d1d669:87761768:86506ed0 name=qy-desktop:15
ARRAY /dev/md/1 metadata=1.2 UUID=e5972077:e26ea9d8:d4fb2d48:0fcaafdf name=qw-desktop:1
ARRAY /dev/md/40 metadata=1.2 UUID=670b6017:b6e63df9:6ac959b2:3b901f51 name=fileserver:40
ARRAY /dev/md/10 metadata=1.2 UUID=9086174e:91b50e01:09a7fda1:eba8b0ad name=qz-desktop:10
ARRAY /dev/md/30 metadata=1.2 UUID=d368acc2:824b3dfa:046c4484:76a1ae7f name=qy-desktop:30
ARRAY /dev/md/50 metadata=1.2 UUID=b96eca9a:ec8726ca:26a39ed7:bb6e7e6c name=fileserver:50
ARRAY /dev/md/30 metadata=1.2 UUID=d368acc2:824b3dfa:046c4484:76a1ae7f name=qy-desktop:30
qy@qy-desktop:~$ sudo nano /etc/mdadm/mdadm.conf
qy@qy-desktop:~$
 
anneranch said:
Would I post a question as such if I have backup? PLEASE -give me some credit.
Click to expand...

I wasn't trying to be a smart ass. You asked about getting to the CONTENT of the drives. If you had performed the first basic step in data management you would have a backup and this would all be moot. Just nuke and pave.

I'm familiar with hardware raid, but no, not software raid so much. In a hardware setup it's generally a hot swapable drive or the raid controller. But again, having a backup in the first place is standard practice. I don't know if 'credit' is due or not. Was trying to help. But thanks for the big middle finger anyway.
 
My question was not about making backup, it was about gaining access to RAID.
I have a backup , but that was NOT why I posted.
So far the question is still unanswered and changing the subject to "backup" is not helping.
I no longer object to changing the subject, it does no good , but to me it s an indicator that the original question was just skimmed over.
 
Making progress.
Run --assemble --scan and manged to activate / access another RAID5.
However , even after deactivating that array and making sure the array I am
after is active
I cannot get past the
Found some drive for an array that is already active: /dev/md/10

qy@qy-desktop:~$ sudo mdadm --assemble --scan

mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Unknown keyword INACTIVE-ARRAY
mdadm: Found some drive for an array that is already active: /dev/md/10
mdadm: giving up.
qy@qy-desktop:~$
 
I haven't done this in years but I believe you need to use the old UUID to make this work. See the article below.

unix.stackexchange.com

mdadm: assembling another array using device removed from active one

In an attempt to enlarge raid1 array on a working machine, I am testing following scenario: remove one of two disks add bigger disk, resync add another bigger disk, resync, grow to raid-devices=3 ...
unix.stackexchange.com unix.stackexchange.com
 
For what its worth I think the old UUID is in the previous output you posted

ARRAY /dev/md/10 metadata=1.2 UUID=3fee7915:97923e1c:5165ee2e:3e5191c5 name=a-desktop:10
 
jpnilson said:
I haven't done this in years but I believe you need to use the old UUID to make this work. See the article below.

unix.stackexchange.com

mdadm: assembling another array using device removed from active one

In an attempt to enlarge raid1 array on a working machine, I am testing following scenario: remove one of two disks add bigger disk, resync add another bigger disk, resync, grow to raid-devices=3 ...
unix.stackexchange.com unix.stackexchange.com
Click to expand...
I have started identical process, and have posted request for help elsewhere.
I am having a problem with "device busy..." and as soon as that is solved I will try what your
article is suggesting.
Appreciate your help very much,
 
jpnilson said:
For what its worth I think the old UUID is in the previous output you posted

ARRAY /dev/md/10 metadata=1.2 UUID=3fee7915:97923e1c:5165ee2e:3e5191c5 name=a-desktop:10
Click to expand...
I am in the process adapting the procedure.
This ID is for the old array, I thing i need ID of one of the old devices being added to new array.
But that is all taking time...
 
Here is the "add" result
qy@qy-desktop:~$ sudo mdadm --assemble /dev/md35 --update=uuid --uuid=b1db014e:d6652122:03547a11:d943e0f0 /dev/sdf24
[sudo] password for qy:

mdadm: /dev/sdf24 is busy - skipping

So it is back to
How to unbusy a device

BUT
the old array is resynching , maybe I need to wait until it is synchronized ??
 
I would certainly let the sync happen. If this is the only device not making it into the array I would create a new file system and add it to the array. If I am understanding what is going on. In theory this should result in no data loss. I will be home in a few hours and will do some research as this is an interesting problem. There must be a log somewhere for this.
 
Status
Not open for further replies.

Members online

Total: 979 (members: 7, guests: 972)

Latest posts

Top