On Wed, 2 Jul 2014 12:04:09 -0400
Lamar Owen <[log in to unmask]> wrote:

> On 07/02/2014 11:40 AM, Andras Horvath wrote:
> > Good thinking would be, but as I mentioned before, I tested this issue on several hardware. With 2 new disks, new enclosures (3 kinds but same type), on different kind of servers and with different USB cables.
> 
> That does seem to cover it quite well; while it is a bit of a long shot 
> it is possible all three enclosures used the same USB to SATA bridge.  
> The odd part is that if it were intrinsic to the kernel or the USB stack 
> of the OS one would think other people would have seen the problem (and 
> thus why you asked the list if anyone else had seen the problem....:-)).
> 
> > I think it must be the software causing this. Especially because the same hardware setup worked fine for long on Debian 6. That's why I turned to the SL mailing list.
> 
> And that's the correct course.
> 
> > So disk is ok, cable end enclosure is ok, and server hw must be too I believe.
> I would tend to agree.  But there has to be something different here, 
> otherwise I would think the list would be inundated with similar 
> reports.  I can't directly test SL 6 in this case, since I don't have an 
> active SL 6 box at the moment, but my CentOS 6.5 laptop and a couple of 
> CentOS 6 servers (Dell PowerEdge 1950) we have here do large USB data 
> transfers frequently without seeing your issue.  It is possible the SL 
> kernel and USB stack have some differences to the CentOS kernel and USB 
> stack, even being built from the same sources; I would not think that 
> likely, and I would think that CentOS 6 in your circumstance should 
> produce the same results.
> 
> In order to actually replicate the results we would need detailed 
> information on the exact hardware that has been used that exhibits the 
> problem (even down to the firmware rev of the WD hard disk), and then 
> someone with that same hardware would need to replicate.  It could be 
> difficult.
> 
> But if you're satisfied with the SATA connection, just use it.  I for 
> one would like to find out what's actually going on, and help figure out 
> how to actually fix it.

I myself would like to tie up all the loose ends, so I think I'll test the very same setup with enclosures from different manufacturers. I'll do this as the first step. I can do this in a couple of weeks when I'll have those and I will report back here with what I've found.

If that doesn't bring up the problem, then I'll test it with different type of disks. I won't test it with different servers because I already tested it with 2 entirely different one. Somehow I believe it is software related but it sure must do something with the external device. I'll see and write back in a couple of weeks.

Andras