Can you explain why copying a folder of 220MB from one HDD to another HDD took 3 hours?


I did a backup, copying from my external HDD whose grub died (see the case Boot EFI stopped working in a removable external HDD on Mac mini) to my laptop's HDD (observe it has a SSD formatted as ext4 with Manjaro). Copying a folder of 220 MB from the external HDD to laptop's HDD took only few minutes, thanks to SSD.

After having reinstalled my external HDD, but copying a folder of 220MB from my laptop's HDD to my external HDD, it took 3 hours! I did not understand.

Observe that external HDD and laptop's HDD are both ext4.

Lots of factors can slow down such an action, to mention some:

  • USB speed (usb2 vs usb3)
  • amount of small files
  • method of copy (cp, rsync etc)
  • drive speeds (rpms)

3 hours for 220MB is way too much. Something isn't right.

I agree. I was just listing some variables. :slight_smile:
I would probably test the health of the receiving hard drive....

1 Like

You mean you reinstalled the system or you reinstalled the HDD into a machine?

You may want to add some details to the bug that most resembles your use case. Some candidates I found on a quick search:

1 Like

Did you put your SSD in a mini drive enclosure with only USB2 support. If so you must use USB cables with dual power USB plugs. If it doesn't have the dual USB on power supply end it will provide inadequate power to your drive.

If it is in a USB 3 enclosure it is best to be plugged in to a USB 3 port.

Inadequate power supply would be my first guess.

Plugging it in to a self powered USB hub is also best.

I know my external HDD is USB 3. But I do not know which my laptop USB generation.

How to know many rpms in both HDDs?

It is what exactly I searched.

Reinstalled the system.

You know very well my laptop model. Is there way to replace USB 2 for USB 3?

My old laptop had an expressport that I bought a USB 3 add in card for. I guess you should make sure there are Linux supported drivers before you buy though.

USB 3 ports are blue.

The rpm's of the drive should be listed on the drives label.

Not always.
My laptop only has USB3 ports and they are all black. :slight_smile:

Really, thanks for that info. I thought that was pretty much a standard.

Yeah, weird 'standard' of sorts (probably early marketing that ended up being less than followed or abandoned altogether later) ... but can confirm just looked at a model that has literally 0 markings for which of the 2 ports is USB2 or USB3 .. kinda guessing based on relation to board etc til I get a sheet.

Good to know, so much for standards?

Ok enough of that silliness. I think I figured out the OP's problem after re-reading his initial post. Correct me if I'm wrong, but I believe he put an ext4 formated drive into an external USB enclosure.

Unless the OS, MOBO, and the controller on the USB enclosure all support garbage collection/trim it won't happen. If you don't trim an SSD regularly then you are going to get worse and worse performance.

You need to make sure you buy an USB enclosure that supports the trim function. There are some in production. Search the forum, I have pointed out controllers that support this function before.

OK, I found my old post:

Me too!

I thought that's a function of the drive's firmware, not the controller. Are you sure about this?

I'm pretty sure. Only really old first Gen SSD's would be likely to not have garbage collection/trim support in their firmware.

I did a fair bit of research on this a while back. The SSD with trim support in the firmware will still not be able to support trimming unless the enclosure has a controller with UASP support.

Even if those are met, the Mobo must also have a UASP capable onboard controller. If his Mobo is old, he still might not have full support. A Mobo produced in the last several years is likely to have a controller with UASP support. The Linux kernel has supported trim over USB for a fair while now. Every link in the chain must be capable of supporting trim over USB for it to work.

You must have the proper hardware for it to work. Read the thread I posted and you'll see what I'm talking about.

1 Like

Indeed! This document shows communication isn't nearly as transparent as I thought it was. I haven't read it all, but some of pictures/diagrams and some definitions show the interface does have an active role. In ex.:

ATA transport protocol layer: Consists of the services and protocols through which application clients communicate with device servers. The ATA transport protocol layer is contained in the host port and the device ports. The ATA transport protocol service interface is defined in this standard in representational terms using ATA transport protocol services (see 4.5). The ATA transport protocol service interface implementation is defined in each ATA transport protocol standard (e.g., ATA8-APT).


If the transport protocol isn't properly implemented on the interface, the affected requests will also fail.

That's good to know. Tanks.

Forum kindly sponsored by