November 2, 2020
Managing storage is crucial to managing servers. This video covers basic storage techniques in RHEL8, including creating partitions using parted, making a file system, runtime mounting, and permanent mounting.
Disks vs Partitions
Let’s start with a quick review of a disk and a partition. The disk is the physical device connected to the motherboard inside the computer (or via an external port). Partitions are how the disk is divided—one disk can have many partitions, but each partition is on a specific disk, at least with basic storage systems.
So, a server will have one or many physical storage devices connected to it. To list all physical devices, run the lsblk command. This demo is performed on a virtual machine, so the disks are vda, vdb, and vdc. This represents three hard disks attached to the virtual machine, simulating three hard disks installed in one server. On a non-virtual computer, you may encounter sda, hda (legacy), or even nveme0n1, a newer type of SSD.
Master Boot Record vs GUID Partition Table
The GUID Partition Table is a much more modern system. Master Boot Record, MBR, is a legacy system that should only be used in exceptional cases but could still be present in older legacy systems. In GPT, up to 128 partitions are available. The default tool to manage partitions in RHEL8 is parted. Issue the parted command followed by whatever physical device name you’d like to manage, so in this case, it would be parted /dev/vdb.
Creating a Partition with parted
Inside parted, the help option lists the available commands and is very helpful. print shows the current partition table, in this case there is none.
So, the first step is to designate a label, either MBR or GPT. We’ll use GPT, so the command is mklabel gpt. By the way, if you wanted MBR, the label would be mklabel msdos, which is counter-intuitive… After you’ve created the label, the easiest way to create a partition using parted is to type mkpart and it will prompt you for the required information. Enter the details as requested, and be sure to print the details before leaving parted to ensure the intended result has been achieved. Then quit parted.
Once you’ve created the partition using parted, issue udevadm settle to ensure that the kernel has picked up the new storage device.
Creating a File System on the Partition
Discussing the various file systems available in RHEL8 would require a separate video; suffice it to say that XFS is the default file system, and we’ll stick with that for this demonstration. To ensure you’re using the correct partition, list them using lsblk. As mentioned, many file systems are available; simply typing mkfs.+TAB+TAB shows a list. In our case, we’ll use the default XFS system, so issue mkfs.xfs /dev/vdb1.
Runtime Mounting
So now we have a partition with a valid file system that is ready to be used by the system. The next step is to create a place where the file system will be mounted – where files saved on the partition will be located within the file system. In our case, we’ll create a new file system called /data, using the mkdir /data command.
Now, let’s mount this file system in runtime to make sure it’s actually working correctly. The command will be mount /dev/vdb1 /data. To double-check this, run mount by itself to find the file system (usually the last entry).
Permanent Mounting
As mentioned, this file system is now temporarily mounted – if the system were to reboot, it would no longer be available. To make it permanent, we need to edit /etc/fstab, where all the file systems are listed. Let’s do that now.
Use vi to open /etc/fstab, and then add a line at the end. First, the device name, the location in the file system where it can be found, the file system type, and then defaults 0 0 will work for this system.
After a quick reboot, listing the mounted file systems using the mount command shows that our partition is permanently mounted and available for use.