Should I replace my SSD?

Submitted ⁨⁨1⁩ ⁨year⁩ ago⁩ by ⁨SexualPolytope@lemmy.sdf.org⁩ to ⁨selfhosted@lemmy.world⁩

cross-posted from: lemmy.sdf.org/post/5332699

I have an SSD that’s around 5 years old now. It used to be in my laptop. But then I upgraded my laptop and put it in a homeserver. It still works perfectly well but from what I’ve read, SSDs fail suddenly without much prior indications.

Do you think I should replace it already? It’s not running any super important stuff. If it dies, it’ll just mean that my media servers will be down for a day, not a super big deal since I have regular backups. I feel bad creating unnecessary e-waste, so I’ll love to know your experience with SSDs and how frequently do you usually replace them.

Also, if you know a tool which can help me detect remaining lifespan of an SSD, that’ll be very helpful. Thanks.

source

Comments

Sort:hotnew top

Fubar91@lemmy.world ⁨1⁩ ⁨year⁩ ago
You can track the health status of most smart enabled ssds. Can use a tool like crystal disk info

Personally i have 2 7 year old ssds going strong without issue. Mainly used for storage and games, so the r/w rates been pretty lower on them.

Ssds do have a total maximum write cycles to nand. Really depends on the use cases over the 5 years.

source
- Appoxo@lemmy.dbzer0.com ⁨1⁩ ⁨year⁩ ago
  Not always does Crystal disk completely shine through the disk.
  Had a sandisk 512GB SSD which was completely fine.
  One day it suddenly became very slow with read and write performance. It was in the <20mb/s range amd painful to recover data from.
  
  Regarding the write cycles: If they ar used up the cells should enter a read only mode so that you should be able to recover the data from. Bad time if it’s the OS though.
  
  source
  - InvertedParallax@lemm.ee ⁨1⁩ ⁨year⁩ ago
    
    Regarding the write cycles: If they ar used up the cells should enter a read only mode so that you should be able to recover the data from. Bad time if it’s the OS though.
    
    This has never happened to me, but I suspect it’s because the controller is the primary failure point here.
    
    source
  - Fubar91@lemmy.world ⁨1⁩ ⁨year⁩ ago
    Agreed, i mainly mention Crystaldisk because its a quick free tool. Definitely reccomend using multiple avenues of info gathering to determine hardware health.
    
    source
    -> View More Comments
- SexualPolytope@lemmy.sdf.org ⁨1⁩ ⁨year⁩ ago
  They’ve not been used too much, I think. My laptop had very typical laptop usage: browsing, reading docs, coding, nothing storage intensive. On the server, the most intensive usage is for PhotoPrism and Jellyfin, and I don’t think that’s anything out of the usual.
  
  source
  - Fubar91@lemmy.world ⁨1⁩ ⁨year⁩ ago
    Id say they are most likely in good health. But anything could happen. Always reccomend having a backup option in place.
    
    source
- ares35@kbin.social ⁨1⁩ ⁨year⁩ ago
  many older ssd are actually better in terms of longevity because slc and mlc typically have/had higher endurance than newer tlc and (especially) qlc.
  
  source
Que@lemmy.world ⁨1⁩ ⁨year⁩ ago
Sounds like you’ve already answered your own question.

It might die. If it does, it’s not a super big deal. You have backups. You don’t like creating unnecessary e-waste.

source
ebits21@lemmy.ca ⁨1⁩ ⁨year⁩ ago
If you have regular backups I wouldn’t worry. If you don’t I worry…. and you should!

source
SheeEttin@lemmy.world ⁨1⁩ ⁨year⁩ ago
They don’t die any more suddenly than a spinning drive, but they do have limited lifespans. Use SMART tools to check the lifespan data. And check for firmware updates since sometimes there are bugs that can cause them to wear out quickly.

But really the important thing is having backups, which you said you have.

source
MangoPenguin@lemmy.blahaj.zone ⁨1⁩ ⁨year⁩ ago
No, no need to replace it unless the wearout in SMART is tripped.

Sudden failure happens in all hardware, keep backups that are tested so you don’t need to worry about it.

source
Dave@lemmy.nz ⁨1⁩ ⁨year⁩ ago
As others have said, keep using it.

But I what I haven’t seen mentioned yet is that you could buy your replacement now. I had an SSD die suddenly (no SMART error) and the most annoying thing was the server being down until a new hard drive could arrive.

source
- macallik@kbin.social ⁨1⁩ ⁨year⁩ ago
  Not a huge deal, but if the SSD goes on to last for X more years, buying an SSD today to save a bit of time will seem pretty poorly thought-out in retrospect
  
  source
  - Dave@lemmy.nz ⁨1⁩ ⁨year⁩ ago
    Make use of it then! Or even better, ZFS mirror or RAID 1 so that if one dies the other holds a perfect copy and can continue running unaffected.
    
    source
dong@lemmy.blahaj.zone ⁨1⁩ ⁨year⁩ ago
Check the life remaining on the drive with crystaldiskinfo it should give you a nice percentage of life but I go off how much data it has seen.

An SSD from 5 years ago is probably rated for around 200TB written, but I would check your model. As long as you are below the TBW with no corrupted sectors, the SSD is fine to keep using. I would probably back it up more frequently, and keep checking crystaldisk for any changes.

source
- Appoxo@lemmy.dbzer0.com ⁨1⁩ ⁨year⁩ ago
  Just do backups.
  
  source
vext01@lemmy.sdf.org ⁨1⁩ ⁨year⁩ ago
Check the output of smartctl and have it run some self-tests.

And as others have said. Backup!

source
Deiskos@lemmy.world ⁨1⁩ ⁨year⁩ ago
Keed using it, set up smartd to check up on it regularly and send you an email if something goes wrong.

source