[lug] External USB 3.0 SATA docks

Sun Nov 18 14:16:34 MST 2012

I need an external dock that will let me plug in a SATA drive to achieve
high performance for large (>500GB+) file transfers, then switch to
another drive, and so on. 

I've attempted this by briding the internal SATA port to an external
connector and connecting that to an Antec MX-1 external USB2/eSATA
enclosure.  It works mostly, but it's never been 100% comfortable.  When
I've used external USB 2.0 docks they've worked well for drives that
come and go, but their speed is now a limitation.

Below is my recent attempt to solve this solution.  If you have
suggestions on working eSATA or USB 3.0 setups, please comment!

I bought a Thermaltake USB 3.0 5G Duet.  It's a two drive external USB
3.0 SATA dock.  Sounds wonderful.  Connected to an inexpensive 1x PCIe 2
port USB 3.0 card.

In real life, this is what I experience:

* With one drive installed, Linux sees the drive and it works.
Initially.

* The drive will disappear periodically - I presume this might be a
power savings related feature.   This causes Linux to lose then
rediscover the drive completely as though I had unplugged it.

* With two drives installed, Linux gets confused.  Palimpsest (Disk
Utility) complains that it has two paths to each drive but multi-pathing
is not configured.

* Palimpsest thinks the bad sector errors being reported on drive A also
apply to drive B.

* I've concluded that the two-drives on one USB 3.0 connection does not
work.  This might be related to a JMicron USB to SATA bridge chip in the
5GDuet which has issues or with Linux support for this kind of setup.

* This may be JMicron port multiplier issue related.

Plenty of googling yesterday and today suggests that there may be issues
with the JMicron chips and that periodic disconnections, even during
large file transfers, is a common problem with several USB 3.0
implementations.  I haven't figured out what the cause is an using an
Amedia PCIe card connected to a JMicron chip based device is one
possibility.  I'm tired and giving up at this point.

Here's an example of the disconnect, re-connect that happens:

Nov 18 13:50:26 elite kernel: [ 5444.756836] usb 9-2: Device not responding to set address.
Nov 18 13:50:26 elite kernel: [ 5445.283746] usb 9-2: Device not responding to set address.
Nov 18 13:50:26 elite kernel: [ 5445.484129] usb 9-2: device not accepting address 10, error -71
Nov 18 13:50:26 elite kernel: [ 5445.535138] usb 9-2: USB disconnect, device number 10
Nov 18 13:50:26 elite kernel: [ 5445.535169] sd 15:0:0:0: Device offlined - not ready after error recovery
Nov 18 13:50:26 elite kernel: [ 5445.541814] xhci_hcd 0000:04:00.0: xHCI xhci_drop_endpoint called with disabled ep ffff8802c11be780
Nov 18 13:50:26 elite kernel: [ 5445.541824] xhci_hcd 0000:04:00.0: xHCI xhci_drop_endpoint called with disabled ep ffff8802c11be7c0
Nov 18 13:50:26 elite kernel: [ 5445.555973] Buffer I/O error on device sdm1, logical block 1289
Nov 18 13:50:26 elite kernel: [ 5445.555976] lost page write due to I/O error on sdm1
Nov 18 13:50:26 elite kernel: [ 5445.555978] JBD2: Error -5 detected when updating journal superblock for sdm1-8.
Nov 18 13:50:31 elite kernel: [ 5450.513639] usb 9-2: new SuperSpeed USB device number 11 using xhci_hcd
Nov 18 13:50:31 elite kernel: [ 5450.524597] usb 9-2: Parent hub missing LPM exit latency info.  Power management will be impacted.
Nov 18 13:50:31 elite kernel: [ 5450.525498] usb 9-2: New USB device found, idVendor=152d, idProduct=0551
Nov 18 13:50:31 elite kernel: [ 5450.525505] usb 9-2: New USB device strings: Mfr=1, Product=2, SerialNumber=5
Nov 18 13:50:31 elite kernel: [ 5450.525512] usb 9-2: Product: USB to ATA/ATAPI Bridge
Nov 18 13:50:31 elite kernel: [ 5450.525517] usb 9-2: Manufacturer: JMicron
Nov 18 13:50:31 elite kernel: [ 5450.525522] usb 9-2: SerialNumber: 1130A5BBFFFF
Nov 18 13:50:31 elite kernel: [ 5450.527863] scsi16 : usb-storage 9-2:1.0
Nov 18 13:50:31 elite mtp-probe: checking bus 9, device 11: "/sys/devices/pci0000:00/0000:00:09.0/0000:04:00.0/usb9/9-2"
Nov 18 13:50:31 elite mtp-probe: bus: 9, device: 11 was not an MTP device
Nov 18 13:50:32 elite kernel: [ 5451.531056] scsi 16:0:0:0: Direct-Access     Hitachi  HDS722020ALA330       PQ: 0 ANSI: 2 CCS
Nov 18 13:50:32 elite kernel: [ 5451.533323] sd 16:0:0:0: Attached scsi generic sg12 type 0
Nov 18 13:50:32 elite kernel: [ 5451.533386] sd 16:0:0:0: [sdm] 3907029168 512-byte logical blocks: (2.00 TB/1.81 TiB)
Nov 18 13:50:32 elite kernel: [ 5451.533945] sd 16:0:0:0: [sdm] Write Protect is off
Nov 18 13:50:32 elite kernel: [ 5451.534515] sd 16:0:0:0: [sdm] No Caching mode page present
Nov 18 13:50:32 elite kernel: [ 5451.534521] sd 16:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 13:50:32 elite kernel: [ 5451.537946] sd 16:0:0:0: [sdm] No Caching mode page present
Nov 18 13:50:32 elite kernel: [ 5451.537956] sd 16:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 13:50:32 elite kernel: [ 5451.583173]  sdm: sdm1
Nov 18 13:50:32 elite kernel: [ 5451.585131] sd 16:0:0:0: [sdm] No Caching mode page present
Nov 18 13:50:32 elite kernel: [ 5451.585142] sd 16:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 13:50:32 elite kernel: [ 5451.585148] sd 16:0:0:0: [sdm] Attached SCSI disk
Nov 18 13:50:34 elite kernel: [ 5452.670692] EXT4-fs (sdm1): mounting ext3 file system using the ext4 subsystem
Nov 18 13:50:34 elite kernel: [ 5452.727090] EXT4-fs (sdm1): warning: maximal mount count reached, running e2fsck is recommended
Nov 18 13:50:34 elite kernel: [ 5452.729140] EXT4-fs (sdm1): recovery complete
Nov 18 13:50:34 elite kernel: [ 5452.729468] EXT4-fs (sdm1): mounted filesystem with ordered data mode. Opts: (null)

Here's what happens when you plug in a second drive:

Nov 18 14:12:30 elite kernel: [ 6767.415218] usb 9-2: USB disconnect, device number 11
Nov 18 14:12:30 elite kernel: [ 6767.440369] Buffer I/O error on device sdm1, logical block 1289
Nov 18 14:12:30 elite kernel: [ 6767.440372] lost page write due to I/O error on sdm1
Nov 18 14:12:30 elite kernel: [ 6767.440374] JBD2: Error -5 detected when updating journal superblock for sdm1-8.
Nov 18 14:12:46 elite kernel: [ 6783.817427] usb 9-2: new SuperSpeed USB device number 12 using xhci_hcd
Nov 18 14:12:46 elite kernel: [ 6783.828447] usb 9-2: Parent hub missing LPM exit latency info.  Power management will be impacted.
Nov 18 14:12:46 elite kernel: [ 6783.829355] usb 9-2: New USB device found, idVendor=152d, idProduct=0551
Nov 18 14:12:46 elite kernel: [ 6783.829364] usb 9-2: New USB device strings: Mfr=1, Product=2, SerialNumber=5
Nov 18 14:12:46 elite kernel: [ 6783.829371] usb 9-2: Product: USB to ATA/ATAPI Bridge
Nov 18 14:12:46 elite kernel: [ 6783.829376] usb 9-2: Manufacturer: JMicron
Nov 18 14:12:46 elite kernel: [ 6783.829381] usb 9-2: SerialNumber: 1130A5BBFFFF
Nov 18 14:12:46 elite kernel: [ 6783.831528] scsi17 : usb-storage 9-2:1.0
Nov 18 14:12:46 elite mtp-probe: checking bus 9, device 12: "/sys/devices/pci0000:00/0000:00:09.0/0000:04:00.0/usb9/9-2"
Nov 18 14:12:46 elite mtp-probe: bus: 9, device: 12 was not an MTP device
Nov 18 14:12:47 elite kernel: [ 6784.835750] scsi 17:0:0:0: Direct-Access     ST325031 0AS                   PQ: 0 ANSI: 2 CCS
Nov 18 14:12:47 elite kernel: [ 6784.836096] scsi 17:0:0:1: Direct-Access     Hitachi  HDS722020ALA330       PQ: 0 ANSI: 2 CCS
Nov 18 14:12:47 elite kernel: [ 6784.837335] sd 17:0:0:0: Attached scsi generic sg12 type 0
Nov 18 14:12:47 elite kernel: [ 6784.837782] sd 17:0:0:0: [sdm] 488397168 512-byte logical blocks: (250 GB/232 GiB)
Nov 18 14:12:47 elite kernel: [ 6784.838017] sd 17:0:0:1: Attached scsi generic sg13 type 0
Nov 18 14:12:47 elite kernel: [ 6784.838328] sd 17:0:0:0: [sdm] Write Protect is off
Nov 18 14:12:47 elite kernel: [ 6784.838910] sd 17:0:0:0: [sdm] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.838917] sd 17:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.839195] sd 17:0:0:1: [sdn] 3907029168 512-byte logical blocks: (2.00 TB/1.81 TiB)
Nov 18 14:12:47 elite kernel: [ 6784.839976] sd 17:0:0:1: [sdn] Write Protect is off
Nov 18 14:12:47 elite kernel: [ 6784.840595] sd 17:0:0:1: [sdn] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.840602] sd 17:0:0:1: [sdn] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.845515] sd 17:0:0:0: [sdm] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.845524] sd 17:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.873282]  sdm:
Nov 18 14:12:47 elite kernel: [ 6784.874379] sd 17:0:0:1: [sdn] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.874390] sd 17:0:0:1: [sdn] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.911480]  sdn: sdn1
Nov 18 14:12:47 elite kernel: [ 6784.912974] sd 17:0:0:0: [sdm] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.912978] sd 17:0:0:0: [sdm] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.912980] sd 17:0:0:0: [sdm] Attached SCSI disk
Nov 18 14:12:47 elite kernel: [ 6784.916514] sd 17:0:0:1: [sdn] No Caching mode page present
Nov 18 14:12:47 elite kernel: [ 6784.916516] sd 17:0:0:1: [sdn] Assuming drive cache: write through
Nov 18 14:12:47 elite kernel: [ 6784.916518] sd 17:0:0:1: [sdn] Attached SCSI disk
Nov 18 14:12:48 elite kernel: [ 6785.559809] EXT4-fs (sdn1): mounting ext3 file system using the ext4 subsystem
Nov 18 14:12:48 elite kernel: [ 6785.605695] EXT4-fs (sdn1): warning: maximal mount count reached, running e2fsck is recommended
Nov 18 14:12:48 elite kernel: [ 6785.607761] EXT4-fs (sdn1): recovery complete
Nov 18 14:12:48 elite kernel: [ 6785.608086] EXT4-fs (sdn1): mounted filesystem with ordered data mode. Opts: (null)
Nov 18 14:12:48 elite kernel: [ 6785.721896] Buffer I/O error on device sdl1, logical block 1289
Nov 18 14:12:48 elite kernel: [ 6785.721900] lost page write due to I/O error on sdl1
Nov 18 14:12:48 elite kernel: [ 6785.721902] JBD2: Error -5 detected when updating journal superblock for sdl1-8.

-- 
Maxwell Spangler
========================================================================
Linux System Administration / Virtualization / Development / Computing
Services
Photography / Graphics Design / Writing
Fort Collins, Colorado
http://www.maxwellspangler.com

-- 
Maxwell Spangler
========================================================================
Linux System Administration / Virtualization / Development / Computing
Services
Photography / Graphics Design / Writing
Fort Collins, Colorado
http://www.maxwellspangler.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.lug.boulder.co.us/pipermail/lug/attachments/20121118/1e8143f4/attachment.html>