Crashing when transfering files

In a couple of days, I'll have another docking for a test from a friend.

It is an Inateck hdd dock with ASM1156 chipset.

Also, the test will be done on a completely new mini-PC with Celeron N5105.

To eliminate all possibilities for the system to crash.

Expectations? Expectations are that the system will be unstable again during SMB transmission on 1Gbit ethernet.

Why? Because this will be already the 4th system completely independent from all the others. And the same problem occurs on all systems.

Why am I doing this and testing it? It simply seems impossible for me to have the same problem with different hardware. And I see that many people use USB docking for disks and it works perfectly for years. Because I suspect that there is a problem somewhere deeper.

Zitat von crashtest

Does this system have one (1) SATA port for a data drive?

This system has 6 SATA ports. The deal is, why are external HDD via USB not working?

Zitat von crashtest

Again, how is your network setup when copying between the source and the destination?
Is a wireless link involved?

Tried wireless, tried ethernet port on MBO directly wired to the router 1Gbps port. Same results.

Zitat von crashtest

Exactly how are you transferring "8GB via SMB"? (Copy and paste, rsync, some other file transfer package? etc.)
Is the same client involved in these transfer tests?
Have you looked at your router?
Have you tried rebooting your router, then running a transfer test?

Copy and paste on win10 machine. Tried different usere, different pc/laptop/phone...

Now problem is even worse! The disk crashes by itself after some time without any file copying. It's like domino effect - once it starts, it doesn't stop.

Yes, I even changed to a router in the meantime (the reason was that the old router didn't have 5Ghz wifi) - now it's actually not a problem either.

I tried rebooting the router, but like I said the drive crashes now even when I don't use SMB.

Let me add. SMB transfer and stability of the disk via the SATA port is quite normal and works properly. The problem is only when the disk is on the usb port - the problem is identical on every system, regardless of docking and independent of the disk.

Zitat von crashtest

Then I would look at interference and, potentially, bad grounds.

interference and, potentially, bad grounds where exactly? Each system used different hardware, in different positions, with different power sources.

Zitat von crashtest

I don't know where you are in the world but:

Most US residential power panels bond neutral (the AC power return) to earth ground (the third round prong on an AC power cord) with a zero resistance connection.
____________________________________________________________

Assuming that neutral and ground are NOT bonded at the panel, or there's an open ground path between the panel and wall sockets, or in the worst case scenario, there's no neutral path and ground itself is the return path to the panel:

In a PC, with a three prong cord, neutral and chassis ground are tied together internally through resistance. So, even with a panel that's not bonded, ground or neutal may not necessary be a problem with your server because all internal components (hard drives included) experience the same voltage state. (Further, the PC chassis and metal box enclosure are close the equivalent of a Faraday cage. External interference is less likely because internal components are shielded.)

However, the external USB enclosure (generally speaking) will be using a two prong plug. With a two prong plug, the ground reference will be derived from the neutral prong. "If" your neutral and ground are not bonded at the power panel, at your house, there may be a difference between the chassis ground of the PC case and the ground as sensed by the external drive enclosure. Differences in ground potential, intermittent grounds and floating grounds can cause all kinds of havoc.

In such a scenario, where there are differences in ground between the PC chassis and the external enclosure, the ground wire of the USB cable itself may be carrying current and, to some degree, a voltage potential. These days, electronics "signals" are already running at very low voltages so having any kind of issue with ground may cause problems (of the inexplicable type you're experiencing).

There's also the "age" of the wiring in the house. NEC standards have changed through the years. With installs that are older, there's the oxidation of copper, loose terminals, etc. Without looking into your panel, measuring ground versus neutral, etc., note that the above is speculation.

It's clear to me that your case is a 1 in a ??? out-layer. You won't find anything exactly like it on the forum.

Alles anzeigen

I am from Europe - Croatia. I can confirm that everything is fine with home installations and grounding.

In Croatia, we use the TT earthing system.

I never had any problems. I work a lot with electronics and microcontrollers because of my profession and I have never had any strange phenomena with electronics due to home installations. This is the first time and it's very strange.

Currently, OMV is running as a VM in Proxmox. I added a 12GB virtual ssd system disk and SMB works without problems. The problem is only the external USB drive. Also before, OMV was installed directly on the hardware and the same problem happened.

There are simply no rules. I have a feeling that the HDD docking chipset simply turns off during transfer or the buffer fails. It's simply not clear to me, it's impossible that every docking has the same error on linux.

Docking works perfectly normally on the Windows (same hardware).

Could this problem possibly be related to UAS?

I did the following:

Disconnect the USB HDD. In a terminal window, run the command sudo dmesg -C.

Now, plug in the SSD and run dmesg with no parameters. The idVendor and idProduct are the two hexadecimal numbers you need to take note of.

nano /etc/modprobe.d/blacklist_uas.conf

At the start of the line of parameters, add the text options usb-storage quirks=aaaa:bbbb:u where aaaa is the idVendor for your device and bbbb is the idProduct. So, with the device above the string will be options usb-storage quirks=2109:0715:u save the file.

after that: update-initramfs -u and reboot.

I will see what happens. For some reason the HDD was stable at least when watching Plex or downloading a torrent. Now when I start Plex the HDD immediately crashes. The situation is getting worse! How is that possible? As if the problem is spreading

How to disable USB Attached Storage (UAS)

Zitat von neoxCRO

Could this problem possibly be related to UAS?

I did the following:

Disconnect the USB HDD. In a terminal window, run the command sudo dmesg -C.

Now, plug in the SSD and run dmesg with no parameters. The idVendor and idProduct are the two hexadecimal numbers you need to take note of.

nano /etc/modprobe.d/blacklist_uas.conf

At the start of the line of parameters, add the text usb-storage.quirks=aaaa:bbbb:u where aaaa is the idVendor for your device and bbbb is the idProduct. So, with the device above the string will be usb-storage.quirks=2109:0715:u save the file.

after that: update-initramfs -u and reboot.

I will see what happens. For some reason the HDD was stable at least when watching Plex or downloading a torrent. Now when I start Plex the HDD immediately crashes. The situation is getting worse! How is that possible? As if the problem is spreading

Alles anzeigen

I'm doing the 4th test SMB transfer of 10gb files right now! For now, everything is stable!

This looks promising so far! I will write everything when I gather more information!

Ok, Ok! This is madness, this is actually madness!

I bought hdd dock/case convinced that it was an ASM1135E Chipset, to be honest I didn't even check.

On the product pages it is clearly stated that it is ASM1135E !!!

But when I look at what lsusb shows: ID 2109:0715 VIA Labs, Inc. VL817 SATA Adapter

WHAAAAT THE ACTUAL F***!?

Well, it's a VL817 SATA adapter and not an ASM1135E!! This is crazy!

First, the problem was the JMicron, and now instead of ASM I run into VIA Labs VL817?

There is no end to this madness. It turns out that VIA Labs has the same flaws as JMicron? Disaster!

Zitat von neoxCRO

I'm doing the 4th test SMB transfer of 10gb files right now! For now, everything is stable!

This looks promising so far! I will write everything when I gather more information!

I moved about 800GB of data back and forth via SMB. The data was in chunks of 50GB to 100GB.

Everything is stable for now, works great! I manage to completely saturate 1Gbps ethernet. Transfer speeds are 100-115MB/s and it is stable!

Zitat von crashtest

If your found or fixed something, write it up. Otherwise, this thread could be confusing for future readers.

After I make sure that the problem is solved, I will make a new post about the problem and the solution. So you can put that post maybe as a sticky on the forum so that everyone can find it.