My first job after college was as a Linux systems administrator at a hosting company.

We offered shared and dedicated hosting on both Linux and Windows servers.

The data center had rows of server racks and stacks of disks, which is where I first learned

how often hard disks can fail.

One evening around 5pm.

I was sitting inside one of our data center cages.

Each data center client had their own secured cage to keep their equipment isolated, and it was my day to be on site.

I was working on a support request for a business client who was having issues with their server.

I SSH'd into the system and started troubleshooting.

After running some tests and making temporary files, I decided to clean up my workspace.

Out of habit, I typed rm-rf star.

At first, nothing seemed unusual.

The command ran for a few seconds, but then I realized something was wrong.

It was taking too long.

My stomach dropped as it hit me.

I wasn't in the right directory.

I had run the command one level too high and was deleting client data.

I had only been at the job for a few months, and this wasn't something I could just undo.

I wasn't pulling a cable or restarting a service.

I was actively erasing files.

Not knowing what else to do, I reached out to my colleagues for help.

Thankfully, we were able to restore everything from backups, and the client never even knew anything had happened.

So, moral of the story, backups, backups, backups.

Just because your data lives in the cloud, doesn't mean a recent college graduate can't accidentally delete it.

That day taught me a valuable lesson.

And ever since, every time I use the rm command, I stop, double check the directory, and take a deep breath before pressing enter.

The piece of a computer that holds everything precious to you, on this episode of In The Show.

The wrong thing to do is just go out and buy a computer, and then learn about it.

You'll learn, but you'll learn a lot of things that maybe you didn't want to learn.

A computer that you buy today will likely be obsolete six months from now, and there's not a dang thing that you can do about it.

My name is Josh, and I'm able to keep this podcast independent and advertisement-free because of support from listeners like you.

If you are finding value in what I'm doing here, consider becoming a paid supporter at members.sideofburritos.com.

And as a thank you, members get early access to new videos, ad-free versions of everything, bonus content, and access to a live monthly Q&A.

Thanks for considering. Now let's get back to the show.

Storage in a computer is where data is kept long-term, even when it is powered off.

This includes your operating system, programs, documents, and games.

But this storage can be slow compared to RAM, which we covered in the last episode.

But RAM is temporary storage that the computer uses only when it's powered on.

When you turn off your PC, data and RAM is lost.

RAM is volatile and very fast, while storage is non-volatile and slower.

Modern personal computers use a few main types of storage drives.

Hard disk drives, HDDs, solid-state drives, SSDs, and non-volatile Memory Express solid-state disks, NVMe SSDs.

An HDD is mechanical.

Inside, it has spinning magnetic platters, called disks, and a moving read-write head, like a tiny record player.

Because of this design, HDDs are physically limited in speed.

The disk can only spin so fast, and the head can only move so quickly.

A typical 3.5-inch desktop hard drive running at 7,200 RPM can deliver around 100-200 MB

per second in sequential data reads and writes, and even slower for random access.

due to seek time. They also come in a 2.5 inch size, commonly used in laptops or external hard

drives. Hard drives are best used for large, inexpensive storage. If you need several terabytes

to store media, backups, or games on a budget, hard drives give the most capacity per dollar.

They are common as secondary drives and desktops. For example, pairing a smaller SSD for the

operating system with a big HDD for files. Also, because they have moving parts, they are more

prone to physical failure and are more sensitive to physical movement. Noise and heat are additional

factors. You can hear them click during use. Solid state drives, or SSDs, have no moving parts.

They use flash memory chips, similar to what's

in a USB thumb drive, but much faster and more sophisticated to store data.

SSDs became popular in the late 2000s and quickly took over as the standard for running

operating systems and software due to their speed. A typical SSD using a SATA interface

can read and write around 500 to 550 megabytes per second, which is several times faster than a hard

drive. But beyond just throughput, SSDs have near instant seek times, meaning accessing small files

scattered around the drive is dramatically quicker. This is why a PC booting from an SSD

feels much quicker, as you are not waiting for a disk to spin. SSDs often come in a 2.5 inch

drive form factor with a SATA connector, making them an easy upgrade for older machines.

There are also M.2 SATA SSDs. These are little gumstick-sized circuit boards that plug into an

M.2 slot, which perform the same as a 2.5-inch SATA drive, just in a different shape.

A SATA SSD is a great all-purpose drive for your operating system, programs, and games.

SSDs are also more shock-resistant, since there are no moving parts, and silence in operation.

The cost per gigabyte has come down a lot, though hard drives are still cheaper for very large capacities.

NVMe, non-volatile memory express, SSDs, are the latest generation of solid-state storage

that connect via the PCI Express bus instead of the older SATA interface.

In practical terms, NVMe drives are...

extremely fast, often 5-7 times faster than the SATA SSDs in sequential throughput,

where SATA tops out around 600MB per second, an NVMe SSD on PCIe 3, which those terms will

cover in a future episode on motherboards, can hit around 3,500MB per second, and newer PCIe 4

models can go up to 7,000MB per second, and even faster on cutting-edge PCIe 5 drives.

These drives typically use the same M.2 form factor, the same gumstick shape, or sometimes

as PCIe expansion cards. NVMe drives achieve this speed by using multiple PCIe lanes and a

streamlined protocol designed for flash, instead of the old ATA-SATA designed for spinning disks.

NVMe SSDs are ideal for performance-critical tasks. If you're doing heavy video editing,

large file transfers, running virtual machines, or just want the fastest boot and application

load times, NVMe is the way to go. As of now, NVMe drives are typically a bit more expensive

per gigabyte than SATA SSDs and require a relatively modern motherboard with M.2 slots

or PCIe slots for an adapter. Also NVMe drives can run hotter due to their speed,

so many come with little heat sinks and it's good to have some airflow over them.

In the past, nearly all desktop and laptop storage was a removable drive. You could swap out the hard

drive or SSD. Nowadays, especially in slim laptops,

people who would be using the new laptop of the SSDs in a можputers' station.

This case is not as good as the limit.

tablets, and some premium devices, storage may be soldered directly to the motherboard.

This means you are unable to remove the storage chips. Now you may be wondering why solder the

storage. For manufacturers, this approach saves space and can be more power efficient. It allows

devices to be thinner and have one less connector that could potentially fail. Apple and some Windows

Ultrabook makers do this to achieve very slim form factors. There's also a minor benefit of

potentially higher speeds if they use multiple chips in parallel, which Apple's recent MacBooks do

this. It also lets the manufacturer prevent the user from upgrading hardware themselves and forcing

them to buy a completely new device. So big corporations love this. Now there's one topic

I've always loved when it comes to storage, and that's RAID.

RAID stands for Redundant Array of Independent Discs.

If you have multiple drives, you can combine them in a RAID setup for better performance and or data safety.

RAID is basically a method to group drives together to act as one unit, using techniques like striping, mirroring, and parity.

This is more common in servers and enthusiast setups than in a typical single PC build,

but if you are building a NAS or just want extra protection or speed with several drives, it's good to know the options.

Before we get into it, RAID is not a backup.

If malware deletes files or you accidentally format data, that deletion is mirrored to the other drives.

RAID only protects against hardware failure, so you would still want external backups for critical data.

Now let's talk about the most common RAID levels. The first is RAID 0 which is striping. This mode

takes two or more drives and stripes the data across them for speed and combined capacity.

For example, if you have two 2TB drives in RAID 0, it will appear as one 4TB drive.

Reads and writes can go roughly two times faster because both disks will work in parallel.

The downside is there is no redundancy at all. If any one drive in the array fails, the entire array

and all data is lost because each file is split between drives. RAID 0 is purely for performance.

With today's super fast SSDs, RAID 0 has lost some of the appeal.

The next one is RAID 1 which is mirroring. This setup requires two drives of the same

size. The system writes identical data to both drives, essentially making an automatic duplicate

of everything in real time. If one drive dies, you have an exact copy on the other, so your data is

safe. The trade-off is you only get the capacity of one drive. Two 2TB drives in RAID 1 give you

2TB of usable storage, since the second drive is a mirror. Performance-wise, you don't gain

write speed because it must write everything twice, but reads can be a bit faster, as it can

read different parts from each disk simultaneously. RAID 1 is a simple way to add fault tolerance

against drive failure. It's popular for boot drives in servers, or anyone who absolutely

can't afford downtime from a single drive failing. The next is RAID 5, which is striping plus

REST

REST

Parity, RAID 5 uses at least three drives and combines them in a way that gives you

both extra speed and one drive fault tolerance.

Data is striped across the drives, like RAID 0, with an added parity block distributed

among the drives.

Parity is a calculated checksum that can rebuild the data of one failed drive.

So any one drive can fail and you won't lose any data, and then you can replace the bad

drive and the array rebuilds itself.

The usable capacity of RAID 5 is N-1 drives.

For example, if you have three 2TB drives in RAID 5, that gives you 4TB of usable storage.

Performance wise, the read speeds are good, often better than a single drive since it can

be read from multiple disks, and write speeds are a bit slower than RAID 5.

zero because of the overhead of calculating parity. RAID 5 is common in NAS devices and

servers for a balanced approach. Many consumer motherboards even support RAID 5 via fake RAID,

or you can set it up in software on Windows and Linux. It's a nice option if you have three or

four drives and want a single large volume with some safety net. Keep in mind, if a drive fails,

the array is running in a degraded state. There's no protection until you replace the drive,

and rebuilding can be intensive, which means that the array can fail during the rebuild.

So if you have a single drive failure and during the rebuild another drive fails,

that means your RAID 5 array will lose all data, which leads us to RAID 6, which is what I use in

my NAS. RAID 6 or RAID Z2.

is dual parity. This is an extension of RAID 5 with two parity blocks instead of one,

which allows for two drives to fail without losing any data.

Standard RAID 6 is often found in enterprise arrays or paranoid home users like myself.

You need at least four drives for RAID 6, since two drives worth of space will be taken up by

parity information. For example, if you have four 2TB drives in RAID 6, this gives you 4TB

of usable storage. With more drives, the efficiency gets better. If you have six 2TB drives in RAID 6,

this gives you 8TB of usable storage. The benefit is higher fault tolerance. You can handle two

simultaneous drive failures, which is reassuring for large arrays or when using not-so-reliable

disks. It also makes array rebuilds less.

risky, as you can have another drive fail during the rebuild and your data will still be safe.

The bottom line on RAID is that for most single PC builders, you will not bother with RAID at all.

You'll use one fast SSD for the OS and maybe a hard drive for data and keep regular backups.

And always weigh the complexity versus benefit, because sometimes a good backup strategy is

better than a complex RAID setup for home use. Now when you format a drive, you choose a file

system, which is the way files are organized on that drive. Different operating systems favor

different file systems. NTFS, or New Technology File System, is the default file system for modern

Windows. If you format a drive in a Windows machine for internal use, you'll likely use NTFS.

It supports large file sizes,

no 4GB limit like older FAT32, large partitions, file permissions and encryption, BitLocker uses

NTFS, compression and journaling, which helps recover from crashes. You'll want to use NTFS

for your Windows C drive and any other internal disks on a Windows PC. Other OSs can usually read

NTFS, but writing to NTFS from, say, macOS requires special drivers. Linux can read and

write to NTFS with proper software, but generally NTFS is best when the drive will be used with

Windows. XFAT, or Extended File Allocation Table, is a file system made by Microsoft as a replacement

for FAT32, introduced in 2006. Its main advantage is broad compatibility. It works on Windows,

macOS and modern Linux.

out of the box, and it supports large files, unlike FAT32's 4GB cap.

XFAT does not have all the features of NTFS. It has no permissions or journaling,

but that's okay for portable storage. XFAT is ideal for formatting external drives or USB sticks

that you want to use on multiple platforms. It's also commonly used on SD cards. Just note that

because it lacks journaling, if the drive is unplugged during a write, there's a high chance

of corruption, so always eject safely. ext4, which is 4th Extended File System, is the standard file

system for Linux distributions. Many use ext4 by default for their main partitions. It's a journaling

file system, reliable and quite performant on Linux. It supports large volumes,

and files, and has features like extents that reduce fragmentation. For Windows users,

ext4 isn't directly usable. Windows doesn't natively read ext4,

so it's mostly for Linux-only environments or Android devices.

The last one I want to cover is ZFS, which stands for Zetabyte File System.

This is kind of a random one, but it's what I use on my NAS. ZFS is an advanced file system,

originally from Sun Microsystems, but it's now open source and popular in server and NAS

communities. It's more than just a file system, it's also a volume manager. ZFS has features like

pooled storage, copy-on-write, built-in RAID, RAID-Z, snapshots, checksumming for data integrity,

and transparent compression. It's designed to

preferable.

rent data corruption, and handle massive amounts of data. Like I mentioned, you'd typically use

ZFS in a NAS or storage server scenario, especially if you want the benefits of RAID Z1,

2, or 3 and self-healing data. TrueNAS, which I use, is a popular free NAS OS, uses ZFS exclusively.

ZFS is great if you have a bunch of drives and want a very resilient storage pool.

It does come with some overhead, it likes having plenty of RAM, and is more complex to administer.

ZFS is worth knowing about, but it's not the go-to for a simple PC storage setup.

Lastly, it's important to mention full disk encryption. This is a security feature,

not a type of drive, but it relates to how you use storage. Full disk encryption means all the data

the drive is encrypted, such that without the correct password or key, it's unreadable.

If someone steals your laptop or drive, encryption protects your personal data

from being accessed by the thief. On Windows, the common solution is BitLocker,

which can encrypt your entire SSD or hard drive. On macOS, there's FileVault. And on Linux,

you have tools like Lux or various encryption options during installation.

When you enable these, you typically enter a passphrase at boot,

or it ties it to your login, and then the drive is decrypted on the fly as you use it.

Full disk encryption can lead to some performance reduction,

but in my opinion, it's a non-negotiable. Take the potential performance hit,

and make sure your data is secure.

In the shulls,

錦

錦

錦

Written, researched, and recorded by me, the Fat32 Podcaster.

If you are listening in an app that lets you rate shows, please take a minute to rate this one.

I would truly appreciate it.

This one time I dared my friend to eat a hard drive.

He asked, how was he supposed to do that?

I told him, one bite at a time.

That's it, take care, and I'll see you next time.