Rob's
/sun/Norco/
Image Archive
I've always wanted the 24
disk supermicro case.
Its loud despite the PWM fans because most MB's bios arn't as agressive as HP's blades.
Has front mounted LEDs
Sleds fit tight with no vibration.
Single backplane with optinal expander.
Over $1k.
so in May 2007 I built my own with three 5in3 trays in a tall
case that could hold 4.
I've never really been happy with the setup.. the drive activity LEDs
are not on the drive trays, but are on the edge of the cage. the trays
are easy to slide in and out, but vibrate, so I use heavy grease to
remove the vibration. Yea, each of the three 5in3 have their own fan,
with two speed settings, and an over temp alarm, but the supermicro 5in3
are better. photos
anyways, Sept 2010 I tried again and ordered a new $400 case.
so that leaves me with:
$250
X7SBE MB two 133MHz PCI-X @ 1,067 MB/s each (PCI Express 4x speed)
$195
Xeon E3110 Wolfdale 3.0GHz 6MB L2 Cache LGA 775 65W Dual-Core CPU
$204
8G ECC ram
$210 two
AOC-SAT2-MV8 PCI-X133 8 port SATA cards
$400
NORCO RPC-4224 case
$78 six
4 Discrete SATA to SFF-8087 (Reverse breakout) Cables
$7
4P in Molex 1 Male to 7 Female Power Extension Splitter Cable
powersupply with at least 6 Molex power connectors
two unused disk slots, but no
expander GUID issues with SATA disks.
If I did it over today, an extra $23 buys one newer stuff and the
ability to add a third 8 port SAS PCIe 4x controller off the south bridge.
$190
X8SIL-F Xeon X3400 / L3400 / Core i3 MB Dual LAN Micro ATX w/ Remote Management
$232
Xeon X3440 Lynnfield 2.53GHz 8MB L3 Cache LGA 1156 95W Quad-Core
$134 * 4 for 16G Registered ECC ram
$244 two
AOC-USAS-L8I PCIe 4x
$90 six
SFF-8087 to SFF-8087 Internal Multilane SAS Cable
Another idea for more zfs arc (system ram) are amd's G34 designs.
$265 H8SGL-F with ether kvm
$280 or
$516 8core single processor
$102 * 4 for 16G Unbuffered ECC ram
$134 * 8 for 32G Registered ECC ram
so what's better four 4+1 or two 8+2 vdevs?
The 4 center fan aren't too loud, sort of like the computer you had 4
years ago. Moving from 12v to 6.3v takes the edge off, but still
noticeable. They won't spin below 5v. But boy, the two fans in the back,
are nuts! In fact you can feel the gyro forces while you hold them.
Dragging them down to 6.1v helps alot, but even then, they are still
very noticeable and won't spin lower than 4.4v. The 1600rpm Scythe
Slipstream is silent at 7.8v like the 1200rpm @ 12v.
The back planes are nice. The strain relief on power and sas connectors
are soldered a second time after the wave bath.
The pleasant surprise was that one can shut off the air on unoccupied
bays and mounting holes for 2.5" disks.
wonder what's going to happen next?
treadstone's proposal is the best for the community, but selfishly
this is a replacement for a tri 5in1 disk theory that I was never happy
with. My goal is to improve on it, not the the stock 4224.
But as a gesture we could do a baseline disk temp graph on the old
box doing a scrub and see how that compares to the modified 4224. In all
cases 120mm fans should offer more CFMs, we just need to balance how
much wattage we pump into the fans :-)
the marketing theory is
4 * 35.3 CFM @ 31 dB(A) (141.2) is less then 3 * 68.54 CFM @ 24 dB(A) (205.6)
and
2 * 57.2 CFM @ 44.5 dB(A) (114.4) compared to 2 * 88.11 CFM @ 33.00 dB(A) (176.22)
yea, two fans in the back... but to treadstone's point.. lets measure
it because we realy don't the effect on deltaP.
A few notes from my build:
The disk powered green LED is perfect brightness.
The disk activity blue LED is fast acting, but dim.
Three good (68.54 CFM @ 24 dB(A) 1200rpm Slipstream
with nothing in the back should be enough for most green disk loads. The
flow out with this setup is quite impressive. I'll graph it to be sure.
Rattles are hard to find and can happen on any metal to metal contact.
Bathroom cawk is our firend.
Fans tightened enough to warp will make a noise.
WD black disks are louder than Slipstream fans.
This build has met its mission of holding more disks, quite, and cooler than the last theory.
Most disk trays have little springs on the sides to stop rattling (my old 5in1 didn't and rattled badly, axle grease was my friend)
Norco "could" rattle because they wiggle in the slot, but I admit I haven't experienced it. Anyway, its easy to add four spring steel loops on the side.
The PS has 5 drive molex feeds, I use them all to maximize copper to backplane.
sweloop64 is correct, graphs from smart, a line for each disk.
I'll be ordering the Norco 120mm midplane, less holes than mine.
the brown/gold thing is a south bridge cooler (xmas gift)
X8SIL-F+x3440
that's a nice setup novadude! I showed you mine, now show me yours :-)
#!/bin/sh
while [ 0 ] ; do
/usr/ucb/echo -n `date +%k:%M`
hdadm display | awk '{
if ($6 == "0K05") printf " " $7
}'
echo
sleep 60
done
| | Old case | New 4224
|
| Rise during scrub | 3° | 2°
|
| Delta T | 6° | ~5°
|
| Max T | 48° | 42°
|
| Resting Max T | 45° | 40°
|
So with 5 120mm 1200rpm Slipstreams there is an overall reduction and
increased control for the 10 black disk. The new green spare is at least
5° cooler than the black drives too. I can hear the air seperate from
the 3 inner fans when I put a cover over the backplane to midplane so
there is a slight deltaP issue. We do have a max operating of 60° we
could decrease the flow and take advantage of the delta from ambient.
naaa.
The first hr is steady state with all five 1200rpm slipstream fans. The
second hr is with the case top covering the mid plane and back plane to
simulate no rear fans. The third hr is with no case top. and the last
two hrs are with the case top replaced. It looks like one needs to run
each state for over 2 hrs, (looking at the scrub graph, perhaps 5hrs)
but it give us an idea on fan options.
Device Serial Vendor Model Rev Temperature
------ ------ ------ ----- ---- -----------
c11t1d0p0 WMATV0109098 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c11t2d0p0 WMATV0086641 ATA WDC WD1001FALS-0 0K05 45 C (113 F)
c11t3d0p0 WMATV0086430 ATA WDC WD1001FALS-0 0K05 46 C (114 F)
c11t4d0p0 WMATV0075456 ATA WDC WD1001FALS-0 0K05 45 C (113 F)
c11t5d0p0 WMATV0078325 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c7t0d0p0 WCAV59169098 ATA WDC WD10EVDS-63U 0A01 33 C (91 F)
c7t1d0p0 WMATV6790152 ATA WDC WD1001FALS-0 0K05 45 C (113 F)
c7t2d0p0 WMATV6789810 ATA WDC WD1001FALS-0 0K05 47 C (116 F)
c7t3d0p0 WMATV6963118 ATA WDC WD1001FALS-0 0K05 47 C (116 F)
c7t4d0p0 WMATV6936555 ATA WDC WD1001FALS-0 0K05 46 C (114 F)
c7t5d0p0 WMATV6770119 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c9t0d0p0 B41BT4D83VBC ATA HDT722525DLA380 A91A 255 C (491 F)
c9t1d0p0 WCAMR2404913 ATA WDC WD3200JD-00K 5J08 0 C (32 F)
c9t2d0p0 WMAP41972847 ATA WDC WD1500AHFD-0 7QR5 38 C (100 F)
c9t4d0p0 WCASY8899432 ATA WDC WD5000AAKS-0 4C07 38 C (100 F)
c9t5d0p0 L592ZJRH ATA Maxtor 7L250S0 1E00 40 C (104 F)
sata1/0
sata1/1::dsk/c9t1d0 Mod: HDT722525DLA380 FRev: V44OA91A SN: VDB41BT4D83VBC
sata1/2::dsk/c9t2d0 Mod: WDC WD3200JD-00KLB0 FRev: 08.05J08 SN: WD-WCAMR2404913
sata1/3
sata1/4
sata1/5
sata2/0::dsk/c7t0d0 Mod: WDC WD1001FALS-00E8B0 FRev: 05.00K05 SN: WD-WMATV6936555
sata2/1::dsk/c7t1d0 Mod: WDC WD1001FALS-00E8B0 FRev: 05.00K05 SN: WD-WMATV6770119
sata2/2::dsk/c7t2d0 Mod: WDC WD1001FALS-00E8B0 FRev: 05.00K05 SN: WD-WMATV6963118
sata2/3
sata2/4::dsk/c7t4d0 Mod: WDC WD1001FALS-00E8B0 FRev: 05.00K05 SN: WD-WMATV6790152
sata2/5::dsk/c7t5d0 Mod: WDC WD1001FALS-00E8B0 FRev: 05.00K05 SN: WD-WMATV6789810
sata2/6
sata2/7
sata3/0::dsk/c8t0d0 Mod: WDC WD1001FALS-00J7B0 FRev: 05.00K05 SN: WD-WMATV0086641
sata3/1::dsk/c8t1d0 Mod: WDC WD1001FALS-00J7B0 FRev: 05.00K05 SN: WD-WMATV0109098
sata3/2::dsk/c8t2d0 Mod: WDC WD1001FALS-00J7B0 FRev: 05.00K05 SN: WD-WMATV0086430
sata3/3
sata3/4::dsk/c8t4d0 Mod: WDC WD1001FALS-00J7B0 FRev: 05.00K05 SN: WD-WMATV0078325
sata3/5::dsk/c8t5d0 Mod: WDC WD1001FALS-00J7B0 FRev: 05.00K05 SN: WD-WMATV0075456
sata3/6
sata3/7
Device Serial Vendor Model Rev Temperature
------ ------ ------ ----- ---- -----------
c7t0d0p0 WMATV6936555 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c7t1d0p0 WMATV6770119 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c7t2d0p0 WMATV6963118 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c7t4d0p0 WMATV6790152 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c7t5d0p0 WMATV6789810 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c8t0d0p0 WMATV0086641 ATA WDC WD1001FALS-0 0K05 44 C (111 F)
c8t1d0p0 WMATV0109098 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c8t2d0p0 WMATV0086430 ATA WDC WD1001FALS-0 0K05 45 C (113 F)
c8t4d0p0 WMATV0078325 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c8t5d0p0 WMATV0075456 ATA WDC WD1001FALS-0 0K05 44 C (111 F)
c9t1d0p0 B41BT4D83VBC ATA HDT722525DLA380 A91A 255 C (491 F)
c9t2d0p0 WCAMR2404913 ATA WDC WD3200JD-00K 5J08 0 C (32 F)
Device Serial Vendor Model Rev Temperature
------ ------ ------ ----- ---- -----------
c11t1d0p0 WMATV0109098 ATA WDC WD1001FALS-0 0K05 39 C (102 F)
c11t2d0p0 WMATV0086641 ATA WDC WD1001FALS-0 0K05 40 C (104 F)
c11t3d0p0 WMATV0086430 ATA WDC WD1001FALS-0 0K05 40 C (104 F)
c11t4d0p0 WMATV0075456 ATA WDC WD1001FALS-0 0K05 39 C (102 F)
c11t5d0p0 WMATV0078325 ATA WDC WD1001FALS-0 0K05 38 C (100 F)
c12t0d0p0 Patriot Memory PMAP None
c7t0d0p0 WCAV59169098 ATA WDC WD10EVDS-63U 0A01 33 C (91 F)
c7t1d0p0 WMATV6790152 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c7t2d0p0 WMATV6789810 ATA WDC WD1001FALS-0 0K05 43 C (109 F)
c7t3d0p0 WMATV6963118 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c7t4d0p0 WMATV6936555 ATA WDC WD1001FALS-0 0K05 42 C (107 F)
c7t5d0p0 WMATV6770119 ATA WDC WD1001FALS-0 0K05 38 C (100 F)
c9t0d0p0 B41BT4D83VBC ATA HDT722525DLA380 A91A 255 C (491 F)
c9t1d0p0 WCAMR2404913 ATA WDC WD3200JD-00K 5J08 0 C (32 F)
c9t2d0p0 WMAP41972847 ATA WDC WD1500AHFD-0 7QR5 34 C (93 F)
c9t3d0p0 WMAPZ0538599 ATA WDC WD3200AAJS-0 1B01 None
c9t4d0p0 WCASY8899432 ATA WDC WD5000AAKS-0 4C07 35 C (95 F)
c9t5d0p0 L592ZJRH ATA Maxtor 7L250S0 1E00 36 C (96 F)
Desired front disk map
| Color
|
| | 0 | 1 | 2 | 3
|
| f white | | | a1 | a2
|
| e yellow | c6 | b6 | b8 | a3
|
| d black | c5 | b5 | b7 | a4
|
| c blue | c4 | b4 | c8 | a5
|
| b green | c3 | b3 | c7 | a6
|
| a red | c2 | b2 | c1 | b1
|
Resulting cabling
| Controler | Cable
|
| c8 | blue2
|
| c7 | green2
|
| c6 | yellow0
|
| c5 | black0
|
| c4 | blue0
|
| c3 | green0
|
| c2 | red0
|
| c1 | red2
|
|
|
| b8 | yellow2
|
| b7 | black2
|
| b6 | yellow1
|
| b5 | black1
|
| b4 | blue1
|
| b3 | green1
|
| b2 | red1
|
| b1 | red3
|
|
|
| a6 | green3
|
| a5 | blue3
|
| a4 | black3
|
| a3 | yellow3
|
| a2 | white3
|
| a1 | white2
|
New desired front map:
A b c D yellow
E f white
a B c D black
E f G h blue
a B c d green
E f G H red
Click on any of the images below for a better view.
The number under the image is the size of the better view in
bytes followed by the date it was last edited and its name.
Click on the name to view the jpeg header.
Click back in your browser to return to this index or the parent of this archive has more images too.
This 338
index.shtml
was served for
(none)
(207.58.244.114) via
(none) on
Tuesday, 02-Nov-2010 20:59:02 EDT using
Wget/1.9.1 last edited by
webservd on
Wednesday, 27-Oct-2010 13:16:46 EDT.
