Dear all, I'm facing issues when rsync-ing files between two hard drives.

Destination drive is contained in an external USB 3.0 enclosure and connected to a bus card installed in a laptop HP ProBook 6450b.

seems to work fine until operation is suddenly aborted by an I/O error. After that, data are no more visible when browsing the drive with Dolphin. Every time it's necessary to unmount and disconnect this external HDD and to connect it again. These sudden disruptions also cause file system to be cleaned with fsck.

After several attempts with several drives, I suspect the bus card is the culprit. Might there be something wrong with some PCIe power saving settings?
Otherwise I can't explain why this bus card, after a certain time, just goes on strike, no matter if data are being transferred over the bus.

Distro is Leap 15.1 with latest kernel version 5.5.x
Thanks in advance!

I had an ExpressCard that was similarly unreliable, I suspected it was overheating, but never fully investigated - I just resorted to using it for regular peripherals and not data transfer.

Do you have external power to the drive enclosure? Many modern small drives draw little enough power during casual use that they can be powered by the bus, even if it wasn't expressly designed for it.

Once you start exercising the drive hard, it would then be pulling too much power through the interface, causing either low-v data errors, or the excess load causing the interface to overheat.

I tried it with both a self-powered external 2.5" and a 3.5" drive with power supply. Bus card can draw additional power from another USB port.

Every usb port has a current limit; on usb-3.0, it's 500mA
I have a usb-3.0 ssd on a usb-3.0 port with no issues. There's no remote power socket, and a 150mm usbA -- usbC cable. That's the way to go if you're buying.

If you have a rotating platter, stepper motors can impart current overloads and general noise. Cabling is important; usb-3.0 depends on a long list of 'if' conditions (Like you can't have two usbA sockets). The longer the cable, the more signal noise.

Update: System booted with option , rsync performed, here's the result:
No broken pipe, no exit code 11, this after several hours of data transfer.

Keep a close eye on cpu temperature with 'acpi=off'. Run 'sensors' regularly.

You seem to have what the police here call 'a definite line of inquiry.' Happy hunting!

CPU fan was during this last test more often loud as well as overall reactivity lower than usual, I guess thermal throttling.
I wonder if there's any way to tweak power management params by using system tools. BTW, I'm going to check out BIOS settings.
May be useful to know, here's the line identifying my USB 3.0 ExpressCard:

What I meant was that if disabling acpi made it work, you have a problem there, which you can probably trace by moving things out of /etc/acpi/events, and see if you can shake the problem that way.