Howdy,

 

I just got done porting over my 8 disk unraid array to a new 24 bay norco build.  The only things that are common between the two nodes are the drives and USB stick, everything else has been completely new.  Start things off, Parity drive is a 5900RPM Hitatchi drive, other drives are a mix and match of whatever was laying around the house.  All drives are pre-cleared before they added to the existing array, no pending sectors were observed after a few pre-clear sessions. 

 

The server was spec'ed out with 16gbs of ram, after some reading it seems the 32bit version of unraid has issues addressing more the 4gb using PAE.  When coping to the array with 16gbs of ram, speed starts at 40MB/s then drops to 1MB/s.  With mem=4095M in the grub options, im getting 22MB/s - 25MB/s Ok performance but expecting 40MB/s-60MB/s.  Parity sync runs at 90MB/s taking around 12 hours to finish.

 

 

Here are some specs on the system:

 

Supermicro 2.0B - X8SAX-O

Xeon 5502 Proc

16GB RAM

Norco 4224

HX 850 PS

 

PCI Cards, all flashed in IT mode with current BIOS:

 

03:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

Subsystem: Super Micro Computer Inc Unknown device 0400

Flags: bus master, fast devsel, latency 0, IRQ 16

I/O ports at 9000

Memory at fb83c000 (64-bit, non-prefetchable)

Memory at fb840000 (64-bit, non-prefetchable)

Expansion ROM at fb880000 [disabled]

Capabilities: [50] Power Management version 3

Capabilities: [68] Express Endpoint, MSI 00

Capabilities: [d0] Vital Product Data

Capabilities: [a8] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-

Capabilities: [c0] MSI-X: Enable+ Mask- TabSize=15

Capabilities: [100] Advanced Error Reporting

Capabilities: [138] Power Budgeting

Capabilities: [150] #10

Capabilities: [190] #0e

Kernel driver in use: mpt2sas

Kernel modules: mpt2sas

 

04:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

Subsystem: LSI Logic / Symbios Logic Unknown device 3020

Flags: bus master, fast devsel, latency 0, IRQ 16

I/O ports at a000

Memory at fb93c000 (64-bit, non-prefetchable)

Memory at fb940000 (64-bit, non-prefetchable)

Expansion ROM at fb980000 [disabled]

Capabilities: [50] Power Management version 3

Capabilities: [68] Express Endpoint, MSI 00

Capabilities: [d0] Vital Product Data

Capabilities: [a8] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-

Capabilities: [c0] MSI-X: Enable+ Mask- TabSize=15

Capabilities: [100] Advanced Error Reporting

Capabilities: [138] Power Budgeting

Capabilities: [150] #10

Capabilities: [190] #0e

Kernel driver in use: mpt2sas

Kernel modules: mpt2sas

 

05:00.0 Serial Attached SCSI controller: LSI Logic / Symbios Logic SAS2008 PCI-Express Fusion-MPT SAS-2 [Falcon] (rev 03)

Subsystem: Dell Unknown device 1f1c

Flags: bus master, fast devsel, latency 0, IRQ 16

I/O ports at b000

Memory at fbab0000 (64-bit, non-prefetchable)

Memory at fbac0000 (64-bit, non-prefetchable)

Expansion ROM at fbb00000 [disabled]

Capabilities: [50] Power Management version 3

Capabilities: [68] Express Endpoint, MSI 00

Capabilities: [d0] Vital Product Data

Capabilities: [a8] Message Signalled Interrupts: Mask- 64bit+ Queue=0/0 Enable-

Capabilities: [c0] MSI-X: Enable+ Mask- TabSize=15

Capabilities: [100] Advanced Error Reporting

Capabilities: [138] Power Budgeting

Kernel driver in use: mpt2sas

Kernel modules: mpt2sas

 

Power Supply is a Corsair HX-850 with 2 Molex per lead hooked up to one backplane. 4 total molex leads from PS to backplanes Using the “in the box” corsair plugs without splitters.

 

          PS -> Molex 4x  -> Backplane 1

-> Backplane 2

                -> Molex 4x  -> Backplane 3

        -> Backplane 4

 

 

System is hooked up via gige to a cisco gige switch, seeing some dropped packets, however old system had the same issue, might be due to 100mbs ASA and homerun HD units on the network.

 

I recently added a Intel 320 SSD for a cache disk, when writing to the cache disk im seeing speeds of 100MB/s.  When writing directly to a disk share im also seeing 20MB/s speeds.

 

 

Thanks,

 

Tim

 

 

syslog-2014-02-04.txt

Hi I almost the exact same specs as you and same issue..find anything?

I have tinkered with the bios settings, nothing has helped yet.  Ive been trying to read up on the new 6.0 build, but the bickering about plugins drown out anything performance wise the new release has.

Hmm I added the Iperf Plugin  and get 1.2 MB/s  this is just running off the USB stick via Unraid..VM in esxi 5.5

 

Iperf on MAC to IPerf on Unraid 5.05 in Esxi vm = 1.2MB/s

 

if I run Iperf MAC to IPERF on Ubuntu VM in ESXI = 80MB/s plus..

 

 

Run a stock bare-metal system for testing. See my sig to disable add-ons. Check for a BIOS update. Use the free version on a spare flash for testing (or backup and then reformat the current flash.). Do some tests using fewer disks. What is the speed with parity unassigned?

All firmwares are current.  Doing a stock speed test today.  I dont have a ton of free time to get another system spun up to see slow disks speeds are, anything else troubleshooting wise i can do with this node?

 

Tim

 

Run a stock bare-metal system for testing. See my sig to disable add-ons. Check for a BIOS update. Use the free version on a spare flash for testing (or backup and then reformat the current flash.). Do some tests using fewer disks. What is the speed with parity unassigned?