usbhd and usbflash: how much difference to flash lifetime?

[bbbhltz]

I've been using Puppy now for a few days at work. Originally I had installed it to a USB with pmedia=usbhd but changed it to pmedia=usbflash.

I understand the difference, but wanted to know from more experienced users regarding how much wear and tear it could cause to a USB flash drive to leave it in usbhd mode?

[Flash]

Here's a related question: has anyone in this forum ever worn out a flash drive any way at all?

[williwaw]

Here's a related question: has anyone in this forum ever worn out a flash drive any way at all?

I asked a similar question in another thread.

I have lost a few USB sticks thru the years, presumably from overuse, but they were cheapies that cost the same as latte at the expresso stand.
For the last 6 months I have been running QV from a PNY USB and I guess it is doing fine. Do you know any way to test for excessive wear or how to monitor the health of a USB?

[bbbhltz]

I'm using a Verbatim brand that I bought just for this because the only ones I had lying around were either old, very limited in size, USB 2.0 or all of the above.

I just search the forum before asking but I guess I need to work on my wording for queries.

I've been very lucky over the past 25 odd years. I've only cooked one HDD, an iPod mini, a single micro SD card and a Sansa Clip Zip.

[rockedge]

Here's a related question: has anyone in this forum ever worn out a flash drive any way at all?

Yes. Two USB flash drives reached the EOL one which I used for a Lucid 5.2.8 that I could use on a public machine in a condominium complex. It was a 1 G flash drive "stick". that started with bad blocks and rapidly got to the point that computers no longer could even recognize it. Formatted as FAT32.

The other was 4 G and the showed the same behavior. Used this one for a portable Firefox and Thunderbird. Both when in use did a lot of read - writes. Also FAT32.

On the other hand I have a 500 G USB hard drive that is 12 years old and still working well. Also using USB adapters on a range of mechanical HDD's scrounged from thrown away computers that come and go like HDD's do.

[wizard]

Here's a related question: has anyone in this forum ever worn out a flash drive any way at all?

Yes, probably over a dozen. Many were no name, but some were name brand, sizes were from 256kb to 32gb. Failures ranged from dropped dead, failure to read, to failure to format.

Recently find that no name and even some name brand drives have terrible read/write speeds.

Currently use Puppy mode 13 and if using Linux ext, it is ext3, to limit the number of writes to the drive.

wizard

[bugnaw333]

Here`s mine...a very old damaged cover and bent 8G Sony USB (EasyOS Kirkstone inside)...I have to cover it with a "gasket maker" to prevent more rusting when holding it.

[Clarity]

Hello @bbbhltz and @Flash

@flash, my experiences is similar to others here. Yes, I have had stick failures and no way to use; thus inability to reformat or store means they landed in the trash.

@bbzhltz I have a comment and a question:
I use DO NOT have any OS installed on a USB stick. In the past, I have installed non-RAM OSes on a USB drive. I am aware of the performance differences of running an OS over a USB port vs via PCI on a motherboard. But, for such, I had little problems in use of those drives.

I dont install OSes on a stick because my experience know that stick's reliability, for me, does not compare to SSD/HDDs. This, is not just me, it is knowledge known in the industry.

BUT (yes I know its a big 'but') for RAM base systems I will use USBs to boot ISO launchers such that since 2019, I use my USBs merely to boot ISO launchers. This is because there is NO writes to the ISO launcher disk, at all, when the OS(s) is running. An ISO boots to desktop from the launcher and no other activity is ever done to the USB following the ISO launch. THUS, I should NEVER have any issue with these 'launch' USBs...ever. This is because the ONLY time the USB is written to, is when it is, either

1. launcher updating (very rare)

2. adding new ISO files

3. deleting old ISO files

I think you can see this use will result in error-free operations much longer than I will live.

In all of my cases, all persistence from one use to another is maintained on a partition of a drive connected to my PC's motherboard. This is because motherboard storage has much more reliable, stable, and yields pretty constant performance expectations.

So for the past 5 years, I have operated trouble-free versus storage use comparisons to all the years prior.

YMMV in whatever use scenario you choose.

[Flash]

As I understand it, write-erase cycles cause flash memory cells to gradually lose their ability to say for sure whether they contain a 1 or a 0 when read. Flash memory manufacturers incorporate "wear leveling" algorithms in their built-in controllers to spread the use around the memory so that all the cells get about the same number of write-erase cycles.

Flash memories also incorporate error correction. As long as not too many memory cells give errors, an error can be corrected and the output of the memory never contains an error. Until the errors become so common that they overwhelm the error correcting algorithm.
As a flash memory "wears out," the number of errors increases until they become so many that the error correction algorithm can't fix them. I think this appears somewhere between 100,000 and a million write-erase cycles for each memory cell. Since the wear leveling algorithm spreads the wear around, the larger the flash drive, the longer it should take for a given cell to reach 100,000 write-erase cycles in a given application.

I believe there's a way to tell when an error has been corrected, but I don't know how. In other words, a flash memory can be tested to see how many errors it is making even though the errors are being corrected.

[dimkr]

ext3, to limit the number of writes to the drive.

AFAIK ext3 is a journaling file system (unlike ext2), which means more frequent writing.

ext4 is also a journaling file system but AFAIK it's a huge improvement compared to ext3 in terms of performance, making ext3 a bad choice unless you want to mount the partition with something like Puppy 4.x.

In addition, it's possible to create an ext4 with journaling disabled, so you get the best of both worlds: less writing (like ext2) but the compatibility, features and performance of ext4. This leaves very little reason to choose ext2 over ext4.

Can you point to documentation that describes advantages of using ext3 on flash media, compared to ext2 and ext4?

[bigpup]

Really, your question is: running in pupmode 13, does it really help to make a USB flash drive last longer?
If limiting the number or writes, has any affect on life time of the drive, it should help to make it last longer.

Ever wonder why you should do backups of data stored on any drives used in a computer?

Any type drive can fail for any reason at any time!

All type drives have a useful life time.

USB flash drives are much better today than they were in the past.

Name brands have more reasons to make reliable ones, than no name ones do.

Mechanical hard drives are really not any more reliable.
Who really uses one any more.
All computers now come with some type of SSD drive for data storage.
Basically a fancy memory flash drive.

A hard drive that no longer was working.

.
.
The only real fact of using data storage drives of any type.

At some time of use, they will fail.

[dimkr]

I have serious doubts about PUPMODE 13's ability to reduce writing to flash drive, due to the way it's implemented in Puppy. Appending 4k to a file writes 4k (normally) but Puppy's save2flash writes the entire file back to the flash drive using cp, which truncates the file and rewrites it in chunks, instead of appending the 4k chunk to the end of the existing contents.

With my typical usage, iostat says that PUPMODE 13 causes more writing and not less.

[stevie pup]

Don't even get be started on the subject of USB sticks. As I've mentioned in another thread somewhere I was trying out Unetbootin with Bookworm64, and at first it wouldn't boot. So I tried again, on a different USB stick, and all fine. But I've used the first stick for other things since then and there's been no problems, so not sure what's going on there.

Even more baffling is another USB stick I have. I can plug it into any Linux system on any laptop and do whatever I want with it, but Windows 7 won't even recognise it's there. I've got others that are same brand and same size, and there's no issues with any of them. Which convinces me even more that USB sticks have a mind of their own.

All my laptops are about 15 years old, all still have the original hard drives, and all of those are still ok.

[wizard]

@stevie pup

This has helped me fix troublesome USB flash drives:

Gparted>Device>Create Partition Table>msdos

Then reformat.

wizard

[wizard]

@Flash

Flash memory manufacturers incorporate "wear leveling" algorithms in their built-in controllers to spread the use around the memory so that all the cells get about the same number of write-erase cycles.

Wear leveling and error correction are common on SSD's and NVME drives, but unfortunately not on consumer grade USB flash drives.

Thanks
wizard

[rockedge]

Who really uses one any more.

Me.

[fredx181]

@stevie pup

I can plug it into any Linux system on any laptop and do whatever I want with it, but Windows 7 won't even recognise it's there

If the USB has a ext2/3/4 (linux) filesystem, Windows cannot recognize it.

[stevie pup]

This has helped me fix troublesome USB flash drives:

Gparted>Device>Create Partition Table>msdos

Then reformat.

Thanks @wizard, but I've done that loads of times with all my USB sticks. Only the odd one that's troublesome for some strange reason.

If the USB has a ext2/3/4 (linux) filesystem, Windows cannot recognize it.

I'm well aware of that, but no, it's FAT32 as are all my USB sticks until I need to reformat for any reason.