Temporary disconnect of HDD in RAID 0 array

alchemist

3/30/24, 2:17 AM

I have a blade server on Rocky Linux 8 with 5 18TB WD Gold HDDs in a software RAID0 built with mdadm. I will accept all the lectures on this and change my ways if you all can help. I needed coffee while I was investigating what appeared to be an empty HD slot (the LED was off) and I popped out one of the drives in the RAID. I quickly put it back in and did a reboot.. to find that I am booting into emergency mode.

Emergency mode

The /var/log/messages file indicates that the NTFS filesystem on the RAID is unavailable, and the system cannot open the RAID component /dev/sde. Checking fdisk also reports that the backup GPT table is not on the end of dev/sdf.

cat /var/log/messages fdisk -l

The output of cat /proc/mdstat shows the following configuration. I recall that I had md126 and md127 mounted. However, md127 shows status inactive. cat /proc/mdstat

When I examine /dev/md127 with mdadm, it gives the following output: mdadm --examine /dev/md127

I haven't taken any steps to resolve the issue as I would like to understand what happened first. I am concerned about losing the data, but not concerned enough to pay for data recovery. I just don't know enough about RAID and particularly software RAID to feel confident taking action. So how bad is this and what is the safest way to proceed?

EDIT: I see the title of this post has been edited by a user from "Temporary disconnect of HDD in RAID 0 array" to "Destroyed RAID 0 array". Is the RAID definitely destroyed? Any insights would be appreciated.

136

1 + 2

raid

mdadm

raid0

Mircea Vutcovici

3/30/24, 3:47 AM

Best way is to create an image of all disks that are part of the array. Usually I save them as an image on a BTRFS or ZFS volume that has ZSTD compression, take a snapshot and then try to recover the data from the snapshoted images.

djdomi

3/30/24, 6:11 AM

it's against all standard ways to run a raid 0 without a backup. no business related environmental will run like this, do basically offtopic. If the data are really important contract some kind of data restore professional

Score:0

Server

alchemist

5/9/24, 1:43 AM

Well, this story has a happy ending!

After using ddrescue to copy all of the 18Tb disks, I used the ntfs equivalent of fsck (ntfsfix) on /dev/md0. The utility was able to restore the filesystem and all of the data has been recovered.

In the first hours of this post, a mod changed the title from "Temporary disconnect of HDD in RAID 0 array" to "Destroyed RAID 0 array and no backups". As it turns out, that was not the case. I have now restored the original title of the post.

+ 0

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Temporary disconnect of HDD in RAID 0 array

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.