On 5/20/19 0:16 , Brian Bennett wrote:
> Is anybody using USB 3.1 mass storage devices with illumos? (FWIW, I'm using 
> SmartOS)
> 
> I've had a problem with several different multi-bay drive enclosures where 
> the enclosure is reported instead of the individual drives. This causes zfs 
> to assign the same devid to each drive, and while I can create a zpool if I 
> try replacing a drive I can't because it says a different device (the last 
> device) is already part of the pool. For example, if I `zpool attach test 
> c1t0d0 c1t0d1` the error message says that c1t0d3 is already part of the pool.
> 
> This happens with multiple manufacturers, and different number of devices in 
> the enclosure. The two main units I have for testing right now are a four-bay 
> QNAP TR-004 and a dual-bay HornetTek Gemini. These devices use a USB-C 
> connector at the device, and a USB-A connection to the computer. Both exhibit 
> the same problem.
> 
> Using the two HornetTek enclosures, for example gives this output form iostat:
> 
> # iostat -En | grep Serial
> Vendor: ASMT     Product: ASM1352R-PM      Revision: 0    Serial No: 
> 000000000000000 
> Vendor: ASMT     Product: ASM1352R-PM      Revision: 0    Serial No: 
> 000000000000000 
> Vendor: ASMT     Product: ASM1352R-PM      Revision: 0    Serial No: 
> 000000000000000 
> Vendor: ASMT     Product: ASM1352R-PM      Revision: 0    Serial No: 
> 000000000000000 
> 
> And from zdb:
> 
> # zdb | grep devid
>                 devid: 'id1,sd@n5000000000000001/a'
>                 devid: 'id1,sd@n5000000000000001/a'
>                 devid: 'id1,sd@n5000000000000001/a'
>                 devid: 'id1,sd@n5000000000000001/a'
> 
> Putting the same disks into a usb 2 enclosure (and using a USB type B 
> connector at the device), is entirely different:
> 
> # iostat -En | grep Serial
> Vendor: HITACHI  Product: HUA723030ALA640  Revision:  Serial No:  
> Vendor: HITACHI  Product: HUA723030ALA640  Revision:  Serial No:  
> 
> And because the serial number is blank rather than a duplicate (though 
> obviously invalid) value, the devids aren't duplicated and the pool behaves 
> properly.
> 
> Has anybody else experienced this? Anybody know what's going on here?
> 

If we're seeing different behavior between using the device as a USB 2.x
and a USB 3.x device there are a few different possibilities that could
be going on that we should at least rule out. What would be useful is to
dump the entire device descriptor tree as a starting point and then see
what's different about what the device is giving us. While I won't be
around for a while to drive this investigation, the simplest way to get
it is in mdb by running:

::prtusb

Which will give a list of devices with indexes then, run ::prtusb -i
<index from above> -v

Robert

------------------------------------------
illumos: illumos-discuss
Permalink: 
https://illumos.topicbox.com/groups/discuss/T400018ace14d8f9b-Me90a52ec4f57f7af96f17c78
Delivery options: https://illumos.topicbox.com/groups/discuss/subscription

Reply via email to