Skip to content
View in the app

A better way to browse. Learn more.

Unraid

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.

learning about HDD specs and have a few questions

Featured Replies

Hi all - I'm in the process of my first NAS build and will be using UnRaid. Long-time PC builder for gaming but relatively light on NAS knowledge, and with that, detailed HDD specs and their meaning/real world implications. As a result I have some HDD related questions to help me get to a final purchasing decision for new drives going forward.

My NAS use cases:

- Primary use for UnRaid will be as a Plex server, I assume from what I read, using a Docker container.

- Secondary use will be for general document storage / private cloud (yes I will have backups)

- Possible use for a private email server

- Other general hobbyist stuff including use of VM's, but not at all a priority

- May play around with RAID and ZFS in the future if I get right into this stuff and see value in moving away from UnRaid - probably 60/40 against this at best in reality though.

 

My Questions:

 

1. Power On Hours: Barracuda Compute power on hour specs are 2400 per year vs 8760 per year for IronWolf NAS and WD Red. Given that 8760 represents an entire year, does this mean that the Barracuda drives can be expected to fail if in an always on state via an UnRaid NAS setup?
 

2. Power On Hours:  Do I count when the drives are in standby/sleep mode, say overnight for example, toward the 2400 hours per year for Barracuda? Or is it only when idling and under load?
 

3. Rotational Speed: As a hobbyist using Unraid or Raid primarily for Plex (and then just general private cloud duties), would there be much of an appreciable difference in system speed/response between 7200rpm drives and 5400rpm? I would rather use 5400rpm for their lower power draw and temps, but don't want to compromise if the speed differential is something most of you consider significant. (I will be using an SSD for write caching with Unraid)
 

4. CMR vs SMR: I have read up about this issue, and feel I have a reasonable understanding, but nonetheless would like community opinions on whether SMR drives would be acceptable for my use cases described above. (Plex notes: Max number of users 10-15, with it being unlikely to have more than 5-6 concurrent users, most common use time likely the evening leaving plenty of idle time for SMR drives to do their thing.)
I only ask about this because here in Australia there are significant price differences between SMR and CMR alternatives, making me wish/hope that it would be ok to go down the SMR route. Cost difference is about 35% so I would really appreciate peoples thoughts on this
(8TB SMR Barracuda Compute @ $245AUD vs 8TB CMR IronWolf or WD Red @ $330AUD)

 

5. Workload Rate: I thought this was an SSD metric, so not sure what it's all about in relation to hard drives? Any insights would be appreciated. It can be as low as 55TB per year for a Barracuda Compute vs 180TB for an IronWolf NAS for example.

 

6. Seagate vs WD: I know there is a historical preference for WD over Seagate in the hobbyist and professional spaces for reliability reasons. Personally I have never had any issue with either.
I wanted to ask whether people still see this reliability factor as a clear differentiator today, or has Seagate pulled up its proverbial socks?

 

Many thanks!

Just a contrarian pint of view, because my server is for my personal use (so i don’t have the budget or really care about drive specs).

 

I literally have a random mix of hard drives. Hitachi, Samsung, WD, Toshiba, Seagate desktop drives, most bought new, some bought used.  WD and Seagate enterprise drives, again mixed new and used. My oldest drives were 35k power on hours when i retired them (and they were desktop drives).

 

Point is, unless you have a service contract and budget, i think worrying about those details isn’t worth it. CMR/SMR makes a difference when writing data. Makes almost no difference when reading. Buy based on $/TB, and don’t worry about the other stuff.  If you’ve got money riding on this, you may have other considerations.

 

 If you want a more authoritative take on this, you can look up the backblaze data. It is pretty eye opening if you’ve never seen it before.

Edited by whipdancer
Clarity.

  • Author
3 hours ago, whipdancer said:

I literally have a random mix of hard drives. Hitachi, Samsung, WD, Toshiba, Seagate desktop drives, most bought new, some bought used.  WD and Seagate enterprise drives, again mixed new and used. My oldest drives were 35k power on hours when i retired them (and they were desktop drives).

 

Point is, unless you have a service contract and budget, i think worrying about those details isn’t worth it.

 

Yeah as it stands I currently have a mix of cheaper desktop drives too, different brands, size, speeds etc. Some are 7+ years old so probably have limited life left, and as it makes sense to get a new, larger (8TB) parity drive now so that I can replace the others over time with similarly larger drives, I figured I should learn about all this and do it properly. 

I have always gone for $/TB in the past but had no concerns about NAS/RAID setups to consider. Interesting to hear you still have that approach in this space. Will be curious to see what others have to say.

On 11/2/2020 at 5:38 AM, nametaken_thisonetoo said:

I have always gone for $/TB in the past but had no concerns about NAS/RAID setups to consider. Interesting to hear you still have that approach in this space. Will be curious to see what others have to say.

 

I generally shuck external WD drives as the cost is ~60% of a new drive bare drive so I get 1 free 'replacement' with every 2 I buy and at least they are new and not refurbished. I've had segates in the past, however they are now mostly shipping shingled drives in the externals there is no price benefit to buying them vs CMR in the larger WD drives.

 

Power on hours, I spin my drives down after 30mins... as they are often idle for long periods so this saves wear and tear. There is always a debate regarding spinning vs startup wear. I've never seen anything conclusive so I save the power. Power on hours is most likely a warranty issue rather than a physical difference, compute HDD are designed for relatively low workloads. They want you to buy a better drive for 24/7 so balance the warranty that way.  Reading an 8TB drive 24 times a year for a parity check soon adds up as wear an tear.

 

WD have been outed for selling 7200 drives labelled as 5400 class. 

You can save a little between different speeds of drives (8TB seem to use the most power for some reason) but if they spin down it's not worth the effort.

The main benefit is cooling, lower power drives are more suitable where there is limited airflow. I initally had a Dell T20 which was fine with 3/4TB drives but an 8TB ran way to warm. 

 

I would avoid shingled as parity due to the write amplificaiton, I don't see the issue in the array for write once, read many data.

 

My logic

In an array with 6 drives, a write to any drive triggers a parity write. The writes will be relatively random across the array drives and every one will required a 'read and rewrite' of a block of tracks on the shingled drive to keep the parity aligned. With 6 drives potentially being updated, thats a lot of extra work. More drives = greater amplificaiton.

 

For an array drive, most writes 'fill up' an empty space so there is less likelihood to need to read and rewrite a block of tracks and even if it does, it's only for the activity on that disk, not across the array.

 

In my view work rate is an increasing issue. Drives are mechanical items built to a price so will always have a finite service life. As drives get larger and tolerances tighter, activities such as partiy checks, rebuilds etc. take longer and longer. I currenly run parity monthly on a 8TB array but even that adds up as above. 2 weekly parity checks on a 18TB drive would be 432 TB/year vs a spec of up to 550 TB/year on the WD Gold.

 

The best advice I have is only store what you need to. 

I recently cleaned out 7TB of old media that had been watched or partially watched. 

The array you build will need to be replaced in 3-5 years so if you throw in $1000 of drives, you're going to be spending that or a good proportion of that every few years.  

 

 

Archived

This topic is now archived and is closed to further replies.

Account

Navigation

Search

Search

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.