Menu
▾
▴
smartmontools-support
|
From: Dennis B. <den...@rw...> - 2010-09-20 18:12:45
Attachments:
smartctl.txt
|
Hi smartmontools-people, i don't know if you are the right people to ask this question, but i try and correct me if i was wrong: I am using a Lian Li EX-50 [1] e-sata case connected to a PCIe 1x SATA controller with SIL 3132 chip. The case can contain up to 5 HDDs and be operated as various RAID levels or as a plain SATA port multiplier. I found several websites that state the case contains a JMicron JMB393 [2] port multiplier, but i have not yet disassembled the case completely to prove it. I am running a Gentoo Linux with an openVZ kernel (based on vanilla 2.6.32.15, openVZ version is budarin.1), i compiled smartmontools from svn-trunk revision 3156. My problem is, that with four of the ports smartctl does show meaningful output incl. error logs and SMART attributes, but in the fifth case it does not. If i number the hdd slots from top to buttom from 1 to 5, it is port 3 that does not give meaningful smart data. This problem occurres in several of this cases with a different hdd types. All hdds work if directly attached to a sata port or if used in another slot as slot 3. I tested e.g. Western Digital WD15EADS-00R6B0 and SAMSUNG HD154UI hdds. I attacked the defect output to this email. If you say this problem is out of your scope and a problem of the hardware vendor, i am ok with it but if you want to find and solve this problem, i can setup a separate linux box with root access for debugging this issue. (dedicated hardware, not openvz). Thank you in advance. Greets, Dennis Birkholz [1] Lian Li EX-50: http://www.lian-li.com/v2/en/product/product06.php?pr_index=331&cl_index=12&sc_index=42&ss_index=115&g=f [2] JMicron SMB393: http://www.jmicron.com/Product_JMB393.htm |
|
From: Tim S. <ti...@se...> - 2010-09-21 10:09:40
|
On 20/09/10 19:12, Dennis Birkholz wrote: > If you say this problem is out of your scope and a problem of the > hardware vendor, i am ok with it but if you want to find and solve this > Maybe the linux-ide would be a better bet? Tim. -- South East Open Source Solutions Limited Registered in England and Wales with company number 06134732. Registered Office: 2 Powell Gardens, Redhill, Surrey, RH1 1TQ VAT number: 900 6633 53 http://seoss.co.uk/ +44-(0)1273-808309 |
|
From: Christian F. <Chr...@t-...> - 2010-09-21 11:34:17
|
Dennis Birkholz wrote: > > I am using a Lian Li EX-50 [1] e-sata case connected to a PCIe 1x SATA > controller with SIL 3132 chip. The case can contain up to 5 HDDs and > be operated as various RAID levels or as a plain SATA port multiplier. > > I found several websites that state the case contains a JMicron JMB393 > [2] port multiplier, but i have not yet disassembled the case > completely to prove it. > According to the specs the Lian e-sata case (and the JMB393) provides RAID support which does not require extra BIOS or driver. This means that its firmware is able to hide the physical disks and simulate a virtual disk. > I am running a Gentoo Linux with an openVZ kernel (based on vanilla > 2.6.32.15, openVZ version is budarin.1), i compiled smartmontools from > svn-trunk revision 3156. > > My problem is, that with four of the ports smartctl does show > meaningful output incl. error logs and SMART attributes, but in the > fifth case it does not. If i number the hdd slots from top to buttom > from 1 to 5, it is port 3 that does not give meaningful smart data. > > This problem occurres in several of this cases with a different hdd > types. All hdds work if directly attached to a sata port or if used in > another slot as slot 3. I tested e.g. Western Digital WD15EADS-00R6B0 > and SAMSUNG HD154UI hdds. I attacked the defect output to this email. > This output shows a disk with all SMART features hidden, except SMART STATUS and one attribute (Temperature). This suggests that the firmware keeps the physical->virtual mapping still active for slot 3 even if RAID mode is disabled. If this is actually the case, this slot may also have a lower I/O performance. Cheers, Christian |
|
From: Martin V. <mar...@gm...> - 2010-09-26 08:35:18
|
Christian Franke <Christian.Franke <at> t-online.de> writes: > > Dennis Birkholz wrote: > > > > I am using a Lian Li EX-50 [1] e-sata case connected to a PCIe 1x SATA > > controller with SIL 3132 chip. The case can contain up to 5 HDDs and > > be operated as various RAID levels or as a plain SATA port multiplier. > > > > I found several websites that state the case contains a JMicron JMB393 > > [2] port multiplier, but i have not yet disassembled the case > > completely to prove it. > > > > According to the specs the Lian e-sata case (and the JMB393) provides > RAID support which does not require extra BIOS or driver. This means > that its firmware is able to hide the physical disks and simulate a > virtual disk. > > > I am running a Gentoo Linux with an openVZ kernel (based on vanilla > > 2.6.32.15, openVZ version is budarin.1), i compiled smartmontools from > > svn-trunk revision 3156. > > > > My problem is, that with four of the ports smartctl does show > > meaningful output incl. error logs and SMART attributes, but in the > > fifth case it does not. If i number the hdd slots from top to buttom > > from 1 to 5, it is port 3 that does not give meaningful smart data. > > > > This problem occurres in several of this cases with a different hdd > > types. All hdds work if directly attached to a sata port or if used in > > another slot as slot 3. I tested e.g. Western Digital WD15EADS-00R6B0 > > and SAMSUNG HD154UI hdds. I attacked the defect output to this email. > > > > This output shows a disk with all SMART features hidden, except SMART > STATUS and one attribute (Temperature). This suggests that the firmware > keeps the physical->virtual mapping still active for slot 3 even if RAID > mode is disabled. If this is actually the case, this slot may also have > a lower I/O performance. I have the same problem with my EX-50. It is indeed a "feature" of the JMB393, because Windows 7 reports the same bogus firmware version (0955) as smartctl. The only program that shows the real SMART data is the JMRaidSetup program (comes on the EX-50 driver disk, works on Linux). According to strace, it sends special commands the the virtualized drive. The RAID mode is disabled with switches on the back of the case. I hope there is a way to configure the JMB393 as a fully-transparent port multiplier. Otherwise, I'll try a cheap SSD in the slot that gets virtualized (I only need 4 bays). Martin Voelkle |
Thanks for helping keep SourceForge clean.
X