36

My netbook has a flash drive instead of a hard disk drive, and I'm using Ubuntu Netbook Remix with ext3 as the file system. I've read some articles concerning flash drive wear, and the main concerns seem to be:

  1. The amount of write cycles - each cell can be written to only a limited amount of times (Wikipedia has numbers ranging from 1,000 to 100,000)
  2. You can only write data on a "sector" once, and after that the whole block needs to be erased to use again - and these blocks are ranging from 16 KB to 128 KB.

These are said to add up so that normal file systems that aren't designed to take this into account end up using wearing out the flash drive by moving small amounts of data.

Now I don't doubt that the problem is theoretically very real. However, I know we tech people get easily carried away by interesting optimization problems, such as designing an alternative file system to combat flash wear. For example it's great to do memory optimization, but if you end up saving 100 KB of memory when there's hundreds of MB available anyway, it's not fixing a real problem.

What I end up getting from all this, is that I shouldn't use normal file systems on flash drives because they quickly eat up the drive. But I'm not convinced. So the question is: Is flash drive wear actually relevant in everyday, normal usage? Is my laptop, using ext3, going to eat up my flash drive in few years... or is all of this rather a theoretical problem that does reduce the usage time, but only by so little that it'll never happen in normal conditions? Or is transparent, hardware wear leveling already being used on netbook flash drives to fix the problem, so that an alternative file system wouldn't even do any good?

Sources: (1), (2), (3), (4), (5), (6)

10 Answers 10

17

Generally speaking, Flash drive wear and tear is always brought up (SSD and USB) However, I haven't seen it.

I have personally found that cheap USB Flash drives for example go faulty and simply do not get recognised well before you actually see wear and tear.

Also, newer drives use technologies that randomise the locations of writes. I suppose, lets say you have a 100 GB drive and fill it up with 99.5 GB's, then you keep using .5 GB over and over again, you can reach the limit, but again, I use SSDs and USB sticks on a daily basis for very heavy use (over the past few years) and generally speaking, the drives die of general failures well before you see this as a problem.

No Experience with alternative File Systems, However I personally wouldn't bother... Use a mature file system and if it fails within a usable time, take it back under warranty. (if in the UK, up to ~6 years under the sale of goods act as you can say it was designed with a fault and not fit for the purpose of storing data... I am not a lawyer, but I took a laptop back 4 years after buying for a similar reason).

Also, for Windows just maybe worth a look in, I remember seeing a product from Diskeeper, that looks interesting - meant to optimise and extend the life of SSD disks, but I am wondering if it is needed and found several articles doubting it (only linked to one) and goes in to detail about wear and tear. Also, I can not longer see the product on their website, so it must of either been scrapped or built into a different edition.

4
  • 1
    You have 6 years to take back an item that was faulty at the time of sale. If manufacturers point out that an SSD has a limited lifetime when they sell it, then you can not return it as faulty later.
    – JamesRyan
    Aug 28, 2009 at 14:20
  • 3
    The main reason for switching to ext 4 from ext 3 on SSD's is the implementation of the TRIM command en.wikipedia.org/wiki/TRIM_%28SSD_command%29 which makes significant differences to performance on drives that support it in firmware.
    – Col
    Aug 28, 2009 at 16:36
  • 8
    So you use your flash drives heavily, they eventually fail, and you don't think this has anything to do with wear?
    – endolith
    Jan 29, 2010 at 15:36
  • Most compact flash devices these days come with wear-leveling built into the controller. Cheaper USB flash drives may not, but most controllers include it now, so most flash devices of all flavors include it, since it's a "free" feature. Aug 16, 2010 at 1:29
9

Installing windows on a compact flash card showed this problem very obviously, killing the card within days under certain typical usage patterns. (Linux is a little easier on them)

SSD drives have wear leveling to extend this to years. If you fill the drive up 90% and then keep making writes it will swap out the files which have remained unchanged in order to extend the flash's lifetime.

Defragmenting does not help on a flash drive because the underlying data is not stored in the pattern that the OS sees. You need to use vendor specific tools.

Flash specific filesystems could extend a drive's lifetime further but at the moment I think this is largely made irrelevant by the progress of drive technology. How many hard disk drives do you actually use that are older than 5 years?

The other point is that when blocks fail, they fail on write so you don't really need to worry about data corruption as with an old magnetic drive that is failing.

So basically as long as your drive has wear levelling it is not really something you need to worry about.

4
  • Killing the card within DAYS? Really? Yeah, granted you're talking about CF card, not an SSD drive there, but it's actually that fast? Wow. Aug 28, 2009 at 18:07
  • 10
    Citation needed. Aug 28, 2009 at 19:48
  • Well my point is that the difference between industrial compact flash and an SSD is only the wear leveling and not the writes per block. And it makes all the difference. It is quite hard to cite relevant estimates since most wear leveling strategies are proprietary and guarded, so any calculation is just hypothetical. In terms of killing plain flash, I have done it personally, it took just 4 days.
    – JamesRyan
    Aug 30, 2009 at 10:57
  • Industrial compact flash usually also has protection (e.g. caps) against unexpected power failed, as if flash is being written during a power fail you can lose the entire block.
    – Michael
    Apr 4, 2013 at 3:37
3

SSD drives use flash based on the 100,000 write cycle technologies, not the 1000. We haven't had flash drives running that long in the real world, but really, except for perhaps the page file on a normal system, the drive isn't getting that many writes. And modern drives do some wear leveling, and automatically compensate for a few bad blocks.

I give the following advice: If you are not doing something insane (a data logger that fills the drive 500 times per second) don't worry about it. Keep good backups, use the system, and in all likelyhood you're going to replace it for faster/bigger parts long before you hit the flash write lifetime.

3

Now that we have been evaluating all the technical aspects of the problem, let's have a more practical approach:

Is flash drive wear a real problem?

The answer is: No, not if you have a reliable backup strategy.

An SSD, like most other computer components, is ultimately bound to fail. The lifetime can be prolonged with a multitude of tweaks, reducing the number of write cycles.

And when it fails you may replace the SSD or more likely buy a new computer because your netbook is most certainly prehistoric by then.

I do have 2 questions myself:

  1. You just got yourself a 300 dollar worth mini computer, how long do you expect it to last?

  2. Why don't people worry as much about platter HDDs as they worry about the wear level of SSDs?

FYI, I still have my wEEE 701 4G, great little road warrior, using it 5 days a week and it's still in top condition ... imagine, after almost 2 years of wear levelling. Portable computers with platter hard drives are no match for those with SSD when it comes to performance and robustness, pat yourself on your shoulder for the great choice you made and get on with it.

2
  • Realistically, I expect the netbook to last at least a few years, I'd give it somewhere from 4 to 10, depending on the usage. Well, in my care, probably just two years, but that's a separate issue. :) But you're getting to the point here - it doesn't matter if ext3 causes some SSD wear if the computer start to break up anyway. But from the articles, it sounds like it's something to really worry about, rather than just a technical curiosity - which leads me to assume it might wear out in a year or two with the wrong filesystem. Aug 28, 2009 at 18:01
  • there is a lot of FUD out there, what size is the SSD? 16 GB? 32? c'mon a replacement will be less than 50 quid, way cheaper than a battery for example which is likely to die long before the SSD. this is a netbook we're not talking about some Uber SSD for a coupla thousand dollars. :)
    – Molly7244
    Aug 28, 2009 at 18:20
2

I am troubled with the same problem; considering a non-journaling filesystem like ext2.
This is a more generic question I am working on.

Summary from the article linked on the first line,

    USB Hard Drives = Ext3 or Ext2  
    USB Flash Drives = Ext2 with “noatime” or “relatime” mount option

I guess USB drives are quite cheap and getting cheaper.
The point is,
would you mind seeing a corruption because of a crash that missed updating your drive data?
or, would you like the speed and (possible) longer drive life?

If you look at ramdisk based distributions like PuppyLinux,
they run off your system memory and sync (maybe) to the disk once in a while.
Gives you speed and (potential) drive life.
That is another trick I am interested in -- Ubuntu on ramdisk.

Meanwhile, I continue to boot a Ubuntu 9.04 on ext4 over a Cruzer drive.
Not too worried about drive life,
but probably about slower performance over a journaling filesystem on USB.
Till we get USB 3.0 all over the place...

1

Manufacturers are also looking at new flash technologies such as NAND Flash which have a much higher number of write cycles. The company I works for uses flash drives on our hardware. Early systems would have a flash failure after 2 years of operation. We quickly replaced the flash with better versions. The current projection is 20+ years in the same application.

1

Firstly, flash wear is not a real problem. Usually something spoils before the lifespan of flash is reached.

Wear levelling is real. It happens in USB drives. But it does not activate until significant wear has occurred. E.G. 1/3 of lifespan. I can't confirm but wear level is probably real in SSD since a small USB drive can implement it.

"Also, newer drives use technologies that randomise the locations of writes. I suppose, lets say you have a 100GB drive and fill it up with 99.5GB's, then you keep using .5GB over and over again,"

This won't work as the 0.5Gb data gets shifted around, thanks to wear levelling. Meaning, storage from the 99.5Gb will be swapped with the original overused 0.5GB storage. That implies the flash cells ages somewhat uniformly.

I shall iterate my point that it is highly unlikely the flash reaches its lifespan before some other failures.

1

I've had a number of those cheap ($10 for 8GB) drives go dead after 1 year due to bad sectors where the log was constantly written.

If you are talking a cheap USB stick drive don't expect it to last that long. Using a non-journal system seems to help, but even system logs that are rotated often still cause wear and failure on those cheap drives.

1
  • I believe he's talking specifically about SSD drives, not USB drives.
    – Taegost
    Jul 30, 2013 at 20:45
0

Answer for 2019. For anything except the smallest and cheapest ssds not any more. In 2018 a hw site tested the actual write endurance of ssds.

The Samsung evo 840 500 GB with 2d tlc flash (the type with the worst endurance unless someone makes 2d qlc) lasted for around 600TB of writes. And the Samsung pro 840 512 GB for 9.1 PB.

Now even the evo line uses 3d tlc flash (they call it vnand) which has more endurance. I don't think anyone has actually measured it but a 1 tb evo 860 or 1 tb evo plus 970 should last for more than 1 PB of writes.

Unless you are trying to destroy them on purpose (and even then it'll take weeks) they will be obsolete long before wearing out.

For usb keys and memory cards (microsd for example) it's still a problem because they are smaller so the writes aren't spread out over that many cells + they don't have any fancy controllers doing wear levelling. Exactly how large of a problem depends on the specific usb key/memory card, use pattern, file system, luck, size, ...

0

This is a very old question, but since people may still read it, I'll bite: I have some NAS servers that I boot from USB flash drives, and when they stop biting, it's always the flash drives.

Sometimes they won't read days at all, sometimes they will be stuck in read only mode. This depends on the drive model and manufacturer.

It's not surprising, because Linux writes log files to the drivers constantly, and so after a few years they probably have logged thousands of writes.
None of the hard drives have failed yet, but hdd life is typically measured in hours of operation, not write cycles.

If you want to use cheap flash memory to run an operating system, a USB SSD (as opposed to flash drive), or a high endurance sd card are good choices.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .