Selecting a hard drive for a Pi4

Within reason ... I don't know how much of a problem it is these days, but a few years ago there were definitely USB chipsets that couldn't handle very large drives.

Try to find out what chipset is being used (the output from lsusb will give you a clue) and Google it.

I have tried 500GB from Kingston in the past 2-3y. I had no issues on USB 2 or 3 on Linux Debian. On the RPi never tried it, I used SanDisk 128GB or

I am sitting here looking at a Samsung SSD and a Samsung 1Tb SSD both T5 wihich mount, write and read happily on a Pi. They were supplied formatted and worked straight away on Windows and Ubuntu. After a bit of fiddling I found that: sudo apt-get install exfat-fuse exfat-utils and mount as an exfat drive worked happily on raspbian no need to format and retained the data interchange use with no problem.

MArtin

If a vendor gave me a line like that, I'd respond with a few choice comments about digital electronic bigotry - and then find another vendor.

I remember the days when Samsung's star was rising. Sadly, they've changed a lot since then.

I wonder sometimes on which planet OPs are living. Why do you think RHEL exists, cause someone has to take the responsibility for the OS and problems and costs related to it. It might be relevant which OS you have, but the described problem might be caused by the OS or whatever the setup is. I am not advocating Samsung, but it is the usual process. Now imagine you are Samsung. You get a call by someone trying to use your product in some exotic configuration. You have to confirm the problem - how are you going to reproduce and confirm it ... no way. OP should have attached the drive to a Windows machine or other supported system and double checked if reproducible. This is usually written on the product description.

After all time costs and the cost must be rectified. This is not only Samsung. In recent years all of the companies in the real economy are having hard time, because the idiots on the stock exchanges put the money in casino economy that is as far away from the real economy as the expectation that someone will help you debug an exotic setup for free.

In message , Deloptes writes

Is Raspbian really that exotic ? If I had been using something that went out with the ark, or was a home grown thing, I could go along with that, but something with who knows how many million installs.

Adrian

I've had drives fail from Maxstor, Seagate, Toshiba, Connor, Western Digital, Hitachi, Xebec & Quantum since 1984.

The only drives that have failed under warranty were Maxstor and Seagate and were replaced FOC when following their claims procedure.

On Fri, 8 May 2020 11:56:25 +0100, Adrian declaimed the following:

To someone in a tech support phone center... Probably...

... but if one had stated it was Debian (8/9/10) I'd expect a better response.

Many years ago I was trying to install interactive Unix on 386 PC chassis. It froze halfway through.

I phoned the UK support desk 'Ah, the BIOS, she is incompatible'. This sounded like BS, so I phoned the parent company in the USA. "Ok where does it freeze..." "well it gets to here..." (sounds of keys tapping madly) "OK, um that's when its looking for a maths co processor, do you have one?" "No", "Do you have the jumper set to say you haven't got one. "I'm not sure" "Well if you haven't it will do that! "Are you sure?" "Yes, I wrote the installation program".

Chalk and cheese really

"The BIOS, she is incompatible", became an in house joke for anything that didn't work...

A useful link. Is there an "inverse" version, i.e., drives known to _not_ use SMR?

The Seagate drive I mentioned earlier

Makes me a little cautious.....

Thanks for posting!

bob prohaska

For some reason the local big-box retailer has WD MyPassport but no "Essentials". It's a 2TB drive, about $15 more than the Seagate. No explicit manufacterer model number, but it might be A-76457056. Again, advertised for backup, with some sort of encryption feature.

Thanks for posting!

bob prohaska

Careful: there is a difference between a device "supporting" an OS from a technical standpoint (i.e. implementing a supported command set) and the company being asked to provide user level support on any random operating system. _They_ don't support the OS, not _it_.

An oddball Linux variant has unspecified errors when connected through an unspecified adapter - they're entitled to say "we don't cover that" - you as a user can't say where the issue is. Yes, they all speak the same protocol (within limits, e.g. SAS differs from SATA) but the question is not at that level. If the question had been "SCT Write Same completes early" then _that_ is a device issue - first line technical support won't have a clue but better support lines will escalate you to the more techncial levels, but at this level you are asking for OS rather than device support.

As an aside I note the Pi has no native ability to connect an SSD so it is presumably connected through a USB adapter of some sort. It hasn't even been mentioned yet alone specified. I'd even have a hunch that is where the problem really was sicne as you note, the rest is well standardised. However, when details like that are missing _I_ wouldn't want to start trying to chase this down.

Den 2020-05-06 kl. 20:53, skrev bob prohaska:

i have this one in a cheap chinese usb2 enclosure on a rpi3+ Model Family: Seagate Momentus 5400.3 Device Model: ST980811AS Serial Number: 5RL073ES Firmware Version: 3.CAE User Capacity: 80,026,361,856 bytes [80.0 GB]

it states Power_On_Hours ... 17035

and it has continiously been on since nov - 2018

it keeps an postgres database which does both read an writes, and daemons on it produces some logs - like 20-100 mb/day

In message , Dennis Lee Bieber writes

I did point out that it was based on Debian, but the answer was still no. I suspect that it was a polite way of filing it under "can't be bothered", but that might be a little too uncharitable.

Adrian

In message , Andrew Smallshaw writes

It was via an adapter, one of these :

But I'm pretty sure that that didn't come up in conversation. It may well be that that was the source of the problem (I think I've suggested that elsewhere in thread).

Adrian

Now imagine how it sounds today. You call the support. It is in Indonesia. They record the question and tell you they will contact you. They you are called from India and being told that it is a known problem with your OS and the hardware, which is built in China. They say because it is too expensive and your OS is not supported you have to return the drive on your expense :)

Sending the drive back was perhaps the right thing OP did.

yeah and who knows what a chipset they build in and what firmware is on.

Sure you did, but it is not that relevant as you returned the drive. It will be relevant with your next drive or you could have tried the Samsung drive with another adapter.

Only the price is astonishing 8.99 ... they save on each part a part of the cent. disgusting! And they never say what chipset it has inside. Some of them even cover the MCs, so that you can not find out.

I recently built UART to RS232 from top quality parts and the most expensive was the box. I now know why they sell boxless for 5,-

regards

'twas ever thus which is why tobacco tins and the like were popular small equipment cases back in the day. I built a radio in a round Ogden's tin, a signal injector in a cigar tube and a clock in a file card box that I can recall.

Fair point.

USB SSDs exist, and Samsung make them. They may have a SATA or NVMe drive inside with an adapter, or that could have a USB interface directly on the controller (like a big USB stick). It is possible the poster has one of those - we don't have enough to go on.

Theo