bunny-lab/docs
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Update Workflows/Linux/Expand XFS Filesystem.md

Browse Source
This commit is contained in:
Nicole Rappe
2025-01-14 17:41:33 -07:00
parent 7c7b25c67a
commit 0ba86dae1d
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

144
Workflows/Linux/Expand XFS Filesystem.md Normal file
View File

@ -0,0 +1,144 @@
**Purpose**:
The purpose of this workflow is to illustrate the process of expanding storage for a RHEL-based Linux server acting as a GuestVM. We want the VM to have more storage space, so this document will go over the steps to expand that usable space.
!!! info "Assumptions"
It is assumed you are using a RHEL variant of linux such as Rocky Linux. This should apply to any version of Linux, but was written in a Rocky Linux 9.4 lab environment.
This document also assumes you did not enable Logical Volume Management (LVM) when deploying your server. If you did, you will need to perform additional LVM-specific steps after increasing the space.
!!! abstract "Oracle Linux Disk / LVM Terminology Idiosyncrasy"
Oracle Linux refers to disks as `/dev/hda` and /dev/hda2` and not something like `/dev/sda` / `/dev/sda2`. You will see certain parts of this document mention `/dev/hda`, in those cases, you may need to switch it to a standard `/dev/sda<#>` in order to make it work in your particular environment.
## Increase GuestVM Virtual Disk Size
This part should be fairly straight-forward. Using whatever hypervisor is running the Linux GuestVM, expand the disk space of the disk to the desired size.
## Extend Partition Table
This step goes over how to increase the usable space of the virtual disk within the GuestVM itself after it was expanded within the hypervisor.
!!! warning "Be Careful"
When you follow these steps, you will be deleting the existing partition and immediately re-creating it. If you do not use the **EXACT SAME** starting sector for the new partition, you will destroy data. Be sure to read every annotation next to each command to fully understand what you are doing.
=== "Using GDISK"
``` sh
sudo dnf install gdisk -y
gdisk /dev/<diskNumber> # (1)
p <ENTER> # (2)
d <ENTER> # (3)
4 <ENTER> # (4)
n <ENTER> # (5)
4 <ENTER> # (6)
<DEFAULT-FIRST-SECTOR-VALUE> (Just press ENTER) # (7)
<DEFAULT-LAST-SECTOR-VALUE> (Just press ENTER) # (8)
<FILESYSTEM-TYPE=8300 (Linux Filesystem)> (Just press ENTER) # (9)
w <ENTER> # (10)
```
??? info "Detailed Command Breakdown"
1. The first command needs you to enter the disk identifier. In most cases, this will likely be the first disk, such as `/dev/sda`. You do not need to indicate a partition number in this step, as you will be asked for one in a later step after identifying all of the partitions on this disk in the next command.
2. This will list all of the partitions on the disk.
3. This will ask you for a partition number to delete. Generally this is the last partition number listed. In the example below, you would type `4` then press ++enter++ to schedule the deletion of the partition.
4. See the previous annotation for details on what entering `4` does in this context.
5. This tells gdisk to create a new partition.
6. This tells gdisk to re-make partition 4 (the one we just deleted in the example).
7. We just want to leave this as the default. In my example, it would look like this:
`First sector (34-2147483614, default = 19826688) or {+-}size{KMGTP}: 19826688`
8. We just want to leave this as the default. In my example, it would look like this:
`Last sector (19826688-2147483614, default = 2147483614) or {+-}size{KMGTP}: 2147483614`
9. Just leave this as-is and press ++enter++ without entering any values. Assuming you are using XFS, as this guide was written for, the default "Linux Filesystem" is what you want for XFS.
10. This will write the changes to the partition table making them reality instead of just staging the changes.
!!! example "Example Output"
```
Command (? for help): p
Disk /dev/sda: 2147483648 sectors, 1024.0 GiB
Model: Virtual Disk
Sector size (logical/physical): 512/4096 bytes
Disk identifier (GUID): 8A5C2469-B07B-42AC-8E57-E756E62D37D1
Partition table holds up to 128 entries
Main partition table begins at sector 2 and ends at sector 33
First usable sector is 34, last usable sector is 2147483614
Partitions will be aligned on 2048-sector boundaries
Total free space is 1073743838 sectors (512.0 GiB)
Number Start (sector) End (sector) Size Code Name
1 2048 1230847 600.0 MiB EF00 EFI System Partition
2 1230848 3327999 1024.0 MiB 8300
3 3328000 19826687 7.9 GiB 8200
4 19826688 1073741790 502.5 GiB 8300 Linux filesystem
```
=== "Using FDISK"
``` sh
pvdisplay # (1)
fdisk /dev/hda # (2)
p <ENTER> # List Partitions
d <ENTER> # Delete a partition
2 <ENTER> # Delete Partition 2 (e.g. /dev/hda2)
n <ENTER> # Make a new Partition
p <ENTER> # Primary Partition Type
Starting Sector: <ENTER> # Use Default Value
Ending Sector: <ENTER> # Use Default Value
w <ENTER> # Commit all queued-up changes and write them to the disk
```
??? info "Detailed Command Breakdown"
1. Use pvdisplay to get the target disk identifier
2. Replace `/dev/hda` with the target disk identifier found in the previous step
**Point of No Return**:
When you press `w` in both cases of `gdisk` or `fdisk`, then ++enter++ the changes will be written to disk, meaning there is no turning back unless you have full GuestVM backups or a snapshot to rollback with, or something like Veeam Backup & Replication. Be certain the first and last sector values are correctly configured before proceeding. (Default values generally are good for this)
## Detect the New Partition Sizes
At this point, the operating system wont detect the changes without a reboot, so we are going to force the operating system to detect them immediately with the following commands to avoid a reboot (if we can avoid it).
``` sh
sudo partprobe /dev/<drive> # Drive Example: /dev/sda (Rocky) or /dev/hda (Oracle Linux)
sudo partx -u /dev/<diskNumber>
```
!!! bug "Partition Size Not Expanded? Reboot."
If you notice the partition still has not expanded to the desired size, you may have no choice but to reboot the server, then re-run the `gdisk` or `fdisk` commands a second time. In my lab environment, it didn't work until I rebooted. This might have been a hiccup on my end, but it's something to keep in mind if you run into the same issue of the size not changing.
``` sh
sudo reboot
```
## Resize the Filesystem
=== "XFS Filesystem"
``` sh
sudo xfs_growfs /
```
=== "Ext4 Filesystem"
``` sh
resize2fs /dev/sda
```
=== "Ext4 Filesystem w/ LVM"
``` sh
# Increase the Physical Volume Group Size
pvdisplay # Check the Current Size of the Physical Volume
pvresize /dev/hda2 # Enlarge the Physical Volume to Fit the New Partition Size
pvdisplay # Validatre the Size of the Physical Volume Increased to the New Size
# Increase the Logical Volume Group Size
lvextend -l +100%FREE /dev/VolGroup00/LogVol00 # Get this from running "lvdisplay" to find the correct Logical Volume Name
# Resize the Filesystem of the Disk to Fit the new Logical Volume
resize2fs /dev/VolGroup00/LogVol00
```
## Validate Storage Expansion
At this point, you can leverage `lsblk` or `df -h` to determine if the usable storage space was successfully increased or not. In this example, you can see that I increased my storage space from 512GB to 1TB.
!!! example "Example Command Output"
Command: `lsblk | grep "sda4"`
```
└─sda4 8:4 0 1014.5G 0 part /
```
Command: `df -h | grep "sda4"`
```
/dev/sda4 1015G 145G 871G 15% /
```