Platter drives worth it anymore?

[jarablue]

I wanted to get a huge 20tb or more platter drive. Just for storing movies, apps and other stuff that I want to keep for awhile. I would transfer them over to my 990pro when needed to be used. Mainly for movies.

Is it worth it to buy platter drives anymore? They are cheap now right? Or just get a ssd? What are some good prices for 20tb and above platter drives with decent performance?

 

[Ardax]

I mean, what else are you going to store 20 TB of stuff on? I mean, sure 16/30 TB SSDs exist, but they're... pricey.

And if you're just playing movies/music off of it, why bother copying it to something else?

 

[jarablue]

Playing movies and music. That's about it. Thought about copying stuff that needed the i/o but I have nothing really. Just want really large storge for that and docs and stuff. I saw the iron wolf pro and exos stuff.

 

[Paladin]

There is one upside to big platter drives, price per TB. The performance is not as good as SSD and if you have one die on you, you lose a lot of data. So you do RAID or backup but then if you have one die on you, the repair/rebuild/restore process takes a long time.

That said, for bulk storage of media, you don't need the high speed of an SSD. You're watching the movie, not rendering/editing it. It's fine if you only get 120-150MB per second performance at peak. Even the most high def video is only like 40-50MB per second, well... 4k blueray can go up to 120MB or so but still well within what a decent spinning hard drive will do. If you encode it to h265 or av1 that bitrate goes way down.

So, if you are ok with the potential loss of a TON of data in a single drive setup or the time for restore or rebuild in a RAID or backup restore scenario, and you are using it for just bulk media storage, then yeah spinning drives are still hard to beat if you care about the price at all.

The alternative is basically a bunch of SSDs at higher price and the complexity can result in other failure modes and duplicate storage, usage problems and other expenses, etc.

 

[w00key]

SSDs have an annoying tendance to just fail one day with zero recovery option available for you.

HDDs tend to make noises first and annouce their impending doom. Not always, but often enough you can use ddrecovery or something to get most of the data off the drive, slowly.

 

[Lord Evermore]

SSDs have an annoying tendance to just fail one day with zero recovery option available for you.

HDDs tend to make noises first and annouce their impending doom. Not always, but often enough you can use ddrecovery or something to get most of the data off the drive, slowly.

Yes, but they are less likely to die overall. You can also monitor them and have some idea of when there is a problem before anything is lost. Once you hear any weird noises from an HDD, some of your data is probably already gone, and monitoring usually doesn't detect anything early.

 

[w00key]

You can also monitor them and have some idea of when there is a problem before anything is lost.

That's the issue. They don't have monitoring info on when they die, often to firmware bugs or corruption in controller metadata like sector remapping index. I never had a SSD fail other than suddenly and catastrophically.

Harddisks have a reallocated sector count and raw read error rates / corrected errors on retry, if any of these increases it's time to retire them. In all my RMAs they were all due to SMART data saying heeey, time to replace it.

 

[steelghost]

If you're thinking of buying a big old bit bucket hard drive, you probably need to budget for a second one and an external USB3 enclosure for it, so you can keep a backup. Despite the discussion about failure modes of SSDs vs spinning drives, either can just decide to stop working at a moment's notice without any warning. So, you need to either take steps to ensure you have a backup, OR accept that you might well lose it one day.

 

[papadage]

Platters still have their uses for mass storage. Streaming media is a perfect use case for them since they can saturate multiple GB network links.

The price per GB is pretty good. I replaced my NAS and populated 6 out of 8 bays in the new one with 18 TB Ironwolves. That leaves me over 62TB of usable space when mapping the volume to Windows.

 

[Lord Evermore]

That's the issue. They don't have monitoring info on when they die, often to firmware bugs or corruption in controller metadata like sector remapping index. I never had a SSD fail other than suddenly and catastrophically.

Harddisks have a reallocated sector count and raw read error rates / corrected errors on retry, if any of these increases it's time to retire them. In all my RMAs they were all due to SMART data saying heeey, time to replace it.

SSDs have those same features in SMART. I've rarely seen them be useful on either type. But with an SSD you can see the wear rate directly with a tool like Crystal DiskInfo and get an idea whether it's degrading. The only time SMART has ever done anything for me with an HDD was when an OEM system would fail to boot because it detected that the drive was so far gone that it was basically already dead and many files were already corrupt beyond recovery. The sudden failure of SSDs is just an alternate way that they can fail compared to hard drives, and in those cases you simply get nothing from them, but overall SSDs have been shown to be more reliable. You're far more likely to have an HDD fail in any of it's many possible ways than for an SSD to fail in any of its ways. The majority of users aren't going to be looking at any monitoring information anyway, and I'm sure you've seen that some of them will happily listen to an HDD grinding away at its own platters while struggling to accomplish the simplest task and think nothing is wrong.

 

[w00key]

Yeah I monitor wear rate and spare area remaining. I haven't seen any that shows anything similar to reallocated sector count, or raw read error rate, both indicate impending doom on HDD.

SSDs go from completly healthy BSOD due to C: or / going AWOL + to not recognized on boot.