Guide update - How to configure and rebuild software RAID #8252

Jessica41 · 2025-08-05T13:07:42Z

What type of Pull Request is this?

Update of existing guide(s)
Fix
Optimization

Description

Updating the guide to include the EFI partition

Mandatory information

This Pull Request shouldn't be merged before: Do not have a date.
This Pull Request content should be replicated for the US OVHcloud documentation : YES

sbraz

Thanks a lot for your work. That's going to be really helpful for customers!
I left a bunch of comments. Sometimes I only commented on one of the two versions but they apply to both.

sbraz · 2025-08-19T12:22:18Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+
+The default RAID level for OVHcloud server installations is RAID 1, which doubles the space taken up by your data, effectively halving the useable disk space.
+
+**This guide explains how to manage and rebuild software RAID after a disk replacement on your server in BIOS mode**


Suggested change

**This guide explains how to manage and rebuild software RAID after a disk replacement on your server in BIOS mode**

**This guide explains how to manage and rebuild software RAID after a disk replacement on your server in legacy boot (BIOS) mode**

Internally, we refer to the boot mode as UEFI or legacy, not BIOS. I think BIOS could be misleading as the UEFI setup menu is still sometimes called "BIOS" or "BIOS setup".

sbraz · 2025-08-19T12:30:33Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+
+For **GPT** partitions, line 6 will display: `Disklabel type: gpt`.
+
+For **MBR** partitions, line 6 will display: `Disklabel type: dos`.


I would not document MBR at all. We've entirely abandoned it for new installations. Only 3% of the installations done in 2024 were in this case and starting from July 2025 it's 0%.

sbraz · 2025-08-19T12:37:47Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+
+For **MBR** partitions, line 6 will display: `Disklabel type: dos`.
+
+Still going by the results of `fdisk -l`, we can see that `/dev/md2` consists of 888.8GB and `/dev/md4` contains 973.5GB. If we were to run the mount command we can also find out the layout of the disk.


You should mention that it's only the case when booted to disk, not in rescue.
Also, mount's output is pretty verbose. I'd just keep lsblk's

sbraz · 2025-08-19T12:40:38Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+To check whether a server runs on BIOS mode or BIOS mode, run the following command:
+
+```sh
+[user@server_ip ~]# [ -d /sys/firmware/efi ] && echo BIOS || echo BIOS


That should be moved at the beginning of the doc when you mention legacy/UEFI.

sbraz · 2025-08-19T12:40:53Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+Alternatively, the `lsblk` command offers a different view of the partitions:
+
+```sh
+lsblk


Why no prompt here? Some other commands have [user@server_ip ~]#

sbraz · 2025-08-19T19:26:44Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+swap : ignored
+```
+
+We enable the swap partition:


Suggested change

We enable the swap partition:

We activate the swap partition:

sbraz · 2025-08-19T19:26:52Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_uefi/guide.en-gb.md

+swap : ignored
+```
+
+We enable the swap partition:


Suggested change

We enable the swap partition:

We activate the swap partition:

sbraz · 2025-08-19T19:27:17Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_uefi/guide.en-gb.md

+swapon /dev/nvme1n1p4
+```
+
+We exit the Chroot environment with `Exit` and unmount all the disks:


Suggested change

We exit the Chroot environment with `Exit` and unmount all the disks:

We exit the chroot environment with `Exit` and unmount all the disks:

sbraz · 2025-08-19T19:27:31Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_bios/guide.en-gb.md

+swapon /dev/sdb4
+```
+
+We exit the Chroot environment with `exit` and unmount all the disks:


Suggested change

We exit the Chroot environment with `exit` and unmount all the disks:

We exit the chroot environment with `exit` and unmount all the disks:

sbraz · 2025-08-19T19:29:10Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_uefi/guide.en-gb.md

+mount /dev/nvme0n1p1 new
+```
+
+Next, we copy the files from the `old` folder to the `new` folder. Depending on your operating system, you will have a similar output. Here we are using debian:


I see in the next subsection where you perform the changes from the OS itself and not the rescue, you run the sync script. Surely it'd work in rescue too, did you try?

@sbraz I did perform all these steps in rescue mode, I just forgot to add the prompt for rescue mode. Doing that in the next update. Thanks!

SlimJ4

Empty bullet point in Case Study 2
Potential consistency issue in "Rebuilding the RAID in normal mode"

SlimJ4 · 2025-10-30T17:25:57Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_uefi/guide.en-gb.md

+**Case study 2** - There have been major system updates (e.g GRUB) to the OS and the ESPs have been synchronised:
+
+- The server is able to boot in normal mode because all the ESPs contain up-to-date information and the RAID rebuild can be carried out in normal mode.
+- 


Either remove the empty bullet point or add a "The server is unable to boot in normal mode..." line in Case Study 2

SlimJ4 · 2025-10-30T19:33:44Z

pages/bare_metal_cloud/dedicated_servers/raid_soft_uefi/guide.en-gb.md

+
+If your server is able to boot in normal mode after a disk replacement, you can proceed with the following steps to rebuild the RAID:
+
+In our example, we replaced the disk **nvme1n1**.


In the previous examples (Disk fail simulation and rescue mode RAID rebuild), we failed the disk nvme0n1, then replaced it and synced it with nvme1n1. They seem to be reversed in the normal mode RAID rebuild example, as this line states "we replaced the disk nvme1n1". This inversion is consistent across the rest of the section.

Jessica41 added 4 commits July 29, 2025 11:06

guide-creation-1

e82f5a7

update

a920302

update

89ff10f

update

611ea4e

Jessica41 marked this pull request as draft August 5, 2025 13:09

Jessica41 added Do not merge yet This Pull Request is awaiting a GO from product teams or has to be merged on a defined date. FIX The Pull Request contains fixes of code or content (typos) labels Aug 5, 2025

Jessica41 and others added 5 commits August 12, 2025 14:44

more updates

15f5d8d

updates

4f1506d

Update guide.en-gb.md

757c129

Update guide.en-gb.md

14f48cb

Update guide.en-gb.md

d258910

Jessica41 requested a review from sbraz August 18, 2025 15:36

sbraz reviewed Aug 19, 2025

View reviewed changes

Jessica41 added 4 commits August 26, 2025 12:02

updates following feedback

cf0c724

Merge branch 'jf-raid' of github.com:ovh/docs into jf-raid

51e1849

updates

9dd56b7

updates

a6b22be

Jessica41 added the Guide creation The Pull Request contains at least 1 new guide (meta.yaml and index edition needed) label Sep 5, 2025

Jessica41 and others added 8 commits September 5, 2025 11:41

meta file creation

0a6c2c6

updates

b0ef94f

Update guide.en-gb.md

2d2c62a

updates

2a191e8

Update guide.en-gb.md

8ef157b

Update guide.en-gb.md

379e720

Update guide.en-gb.md

235be5f

create FR and minor fix

c063b2b

SlimJ4 requested changes Oct 30, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Guide update - How to configure and rebuild software RAID #8252

Guide update - How to configure and rebuild software RAID #8252

Uh oh!

Jessica41 commented Aug 5, 2025

sbraz left a comment

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

sbraz Aug 19, 2025

Jessica41 Aug 25, 2025

SlimJ4 left a comment

SlimJ4 Oct 30, 2025 •

edited

Loading

SlimJ4 Oct 30, 2025 •

edited

Loading

Labels

4 participants


		The default RAID level for OVHcloud server installations is RAID 1, which doubles the space taken up by your data, effectively halving the useable disk space.

		This guide explains how to manage and rebuild software RAID after a disk replacement on your server in BIOS mode


		For GPT partitions, line 6 will display: `Disklabel type: gpt`.

		For MBR partitions, line 6 will display: `Disklabel type: dos`.


		For MBR partitions, line 6 will display: `Disklabel type: dos`.

		Still going by the results of `fdisk -l`, we can see that `/dev/md2` consists of 888.8GB and `/dev/md4` contains 973.5GB. If we were to run the mount command we can also find out the layout of the disk.

	We enable the swap partition:
	We activate the swap partition:

	We exit the Chroot environment with `Exit` and unmount all the disks:
	We exit the chroot environment with `Exit` and unmount all the disks:


		If your server is able to boot in normal mode after a disk replacement, you can proceed with the following steps to rebuild the RAID:

		In our example, we replaced the disk nvme1n1.

Guide update - How to configure and rebuild software RAID #8252

Are you sure you want to change the base?

Guide update - How to configure and rebuild software RAID #8252

Uh oh!

Conversation

Jessica41 commented Aug 5, 2025

What type of Pull Request is this?

Description

Mandatory information

sbraz left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

SlimJ4 left a comment

Choose a reason for hiding this comment

SlimJ4 Oct 30, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

SlimJ4 Oct 30, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Labels

4 participants

SlimJ4 Oct 30, 2025 •

edited

Loading

SlimJ4 Oct 30, 2025 •

edited

Loading