Are You Looking For A Less Expensive Norco 4220 / 4224 Alternative? - Page 7

No worries.

Just a short delay.

BTW, for those who are going to need to order these -- don't order them from startech directly. For me to get 2 15-packs, they were going to charge more for UPS ground shipping ($16) than they were charging for the items themselves (their next-day shipping cost was $156 ).

They have them on amazon for $5 per 15-pack with free shipping. Just search for Startech STANDOFF632.
Edited by NewOrlnsDukie - 6/23/12 at 3:41pm

I'd plan for 10w per drive, plus whatever the rest of your system will need. I'm not sure you're going to be able to get a pair of ATX power supplies in that case, and I'd suspect you'll need special hardware to handle loud balancing and fail over switching as well. Not sure if anything like that exists (it probably does, somewhere out there) if you did manage it. Sounds like a headache. You can get single redundant ATX power supplies, you'd have to power down the system and manually install a new PSU though. There may be other options that I'd not aware of.

Anyway, I got my system on Friday. Unfortunately it came a bit banged up - both of the front handle brackets are bent, and the handle on the left side broke off as I was lifting the case out of the box. This side (left) was bent worse than the others and bent forward as to prevent the drives from being removed. I ought to be able to bend the brackets straight and the reattach handle, so no big deal. I got a scratch and dent server for the scratch and dent price!

Still haven't figured out what I'll be able to do wrt virtualization on this hardware - I spent a little time looking up raw device mapping for ESXi and Xen, and it seems that it wouldn't work in this scenario since the drives would be connected to the HBA cards, which are PCI-X devices and need VT-d(or whatever the AMD equivalent is). I may just go with a plain jane WHS server after all.
Edited by mesasone - 6/24/12 at 7:01am

Does anyone know where to purchase Rails for this Chassis? Thanks for your help.

My thought is the host OS will handle the drives directly and then have the VM guests access them as virtual HDs (or iSCSI blocks). This way I avoid the VT-d issue.

I looked around and found what I think is a compatible set (from AIC), but I can't find anywhere that sells them. You can probably just measure the length of the case and buy a set of generic rails that will accommodate the case. Or, do what I did and just use a vented server rack shelf and sit the case on it, then secure it to the front of the server rack.

The slides alone are going to be hard to find. Somebody,somewhere recycled the racks that held all these servers but probably didn't pull the slides at the same time. No doubt the slides are long gone.

So you're faced with having to buy a whole rail/slide set. I picked up two sets via ebay. Search for "FRR-FR28-TW-R". They're made by AIC and are 28" deep.

Otherwise you could always pick up any other sets that would handle the weight load. They're using two sets of cheap solid slides, versus much more expensive ball bearing slides. These boxes, when fully loaded, are going to be HEAVY, so you need to use decent ones. I've racked many a machine over the years and you cannot go wrong with buying good quality rail sets. In many situations I've just drilled new mounting holes in either a case or the rail when they don't match up. Sure, it's convenient having a threaded hole in the case, but it's not that much more work to just use a nut and bolt (just be sure the head of the bolt clears the gap in-between the rail and slide).

I strongly urge anyone thinking of mounting one of these beasts to NOT USE SHELVES. You will regret it if/when you need to service something inside the box and have to pull the whole thing out. Buck up and spend the money on appropriate rails. Your back will thank you for not injuring it later, when you've long since forgotten how heavy this eff'ing thing was to start with, and is now even heavier will all those drives loaded...

I've also found it helpful to add cable management arms. This way the cabling can stay attached and out of the way when you slide the machine out. Any generic arm can be made to work so don't fixate on finding something perfect. Most of the time I just use a set of zip ties on the power cord to secure it to the chassis. Run the power cord close to the case and FIRMLY zip tie it there. Then FIRMLY zip tie the cord to the closest end of the cable arm. Yes, you'll be pulling on the power cord when you slide out the machine, but it's not like there's a ton of pressure on it.

I've also found using 14' is the ideal length of cords for ethernet and KVM connections. Put your network and KVM switch in the middle/rear of the rack. Then you'll have enough length to handle the rack slide being extended and still be able to get up to the switches. For machines right above/below the switches you could use a shorter length, of course, but 14' still ends up being good. Not too short, not too long.

Thank you all for your help. / I am a newbie to this forum. so any response is greatly appreciated! Wkearney99 thank you for you post i found the rails you spoke of. very much appreciated.

and to the rest thank you as well. look forward to further discussions.

Anybody have any luck with the Sata cards in ESXi 5? I've searched around and it seems there is a hacked driver for esxi 4.1, but nothing for 5.

I was never able to get those cards to work with ESXi 5, and the 4.1 driver hack never worked properly, either, unfortunately. It seems the only option for virtualization with the hardware as shipped is Hyper-V.

If the cards have Linux drivers, Xen could be an option. Xen supports PCI-passthrough without VT-d to a paravirtualized guest. But it is limited to guest OSes (DomU) that supports a Xen-aware kernel for it to work.

Hey Litlgi74 - I noticed you don't have anything plugged into the 4-pin +12V connector. Have you had any issues with this, or is it just unnecessary?

I actually ended up ordering this PSU for my upgrade of this server, as it has all the necessary connections for the motherboard. It may not be necessary unless you're running dual procs and maxing out the available slots on the board, though.

https://www.superbiiz.com/detail.php?name=PS-SS650AM

Same here. I removed both brackets and bent the case back enough that all drive bays work now. Does anyone know if you can purchase replacement brackets that would work?

My second server had bent handles--one was just slight but the other one was a big bend.

I just took the bracket out to the shop and hammered it back into shape (They are thick metal--they can take a rehammering back into shape).

Picked up a pair of E5450's for $94 and added some passive heatsinks.

Supermicro 1U copper heatsink: http://www.ebay.com/itm/280888242425




Arctic Silver, placed a 120mm fan over the pair pulling air away from the heatsinks. Idle's around 35°C, full load around 65°C so it seems to be doing pretty well like this:



Getting rid of the 60mm screamers made a huge difference, but ultimately the PSU fans are too obnoxious for me. Ordered the fan wall, back panel, three 120mm and two 80mm fans and a Cooler Master PSU with a 135mm fan. Should quiet things down a good bit and also increase airflow.

For those thinking of upgrading, Newegg has 15% off server motherboards until 7/5.
Code: EMCNDHA25

I've been thinking about getting http://www.newegg.com/Product/Product.aspx?Item=N82E16813182348
$399 after coupon.

They would put those on sale a week after I ordered from elsewhere

That's a great board, btw. One thing to consider is that if you're interested in using virtualization, the onboard SAS controller apparently cannot be passed through directly to a VM due to the way it shares the PCIe bus. But if you're not planning on using ESXi or Xen, you're probably fine.

Got my rails for the server today--pretty nice!

FRR-FR28-TW-R are the correct ones and you need 2 sets.

If youve got a kill-a-watt it'd be interesting to know what youre seeing for inital boot power consumption, and thne with a number of spindles running. I may ditch the 3x redundant supplies if consumption can be driven lower with a single supply. That and im curious what those nw xeons draw

I've got two Xeon E5430s and I'm using two of the three power supplies (oddly enough, using just one of them drew more power than using two). I'm doing a RAID migration (added two new drives in RAID 10), so all 6 of my drives are spinning, and I'm sitting at about 175 watts.

Has anyone found the cables needed for the LED status indication on the backplane? I talked to AICIPC and found that the part number of the cable I need is CBL-LED-TW-02, but they don't sell them - I was told I have to talk to the supplier because the cables usually come with the chassis. So far I've emailed a few RSC-4ED2 suppliers but I haven't heard anything back from them. Anyone else have any luck finding these cables?

I have the status indicator (blue light), but no red lights that light up when there's a fault. I also don't use the Supermicro RAID cards, either - I use 3Ware 9500S cards. The Supermicro cards send the status info over the special Supermicro SATA cables, but 3Ware doesn't use them. I'm not having any luck connecting the backplane to the card. I'm guessing this is a Supermicro SATA backplane - does anyone know the part number? I might be able to find a conversion process that require additional cards.
Edited by kegobeer - 6/30/12 at 9:20am

After looking at the chassis, I'm pretty sure it's not a Supermicro backplane - that would be odd considering the chassis and SATA bays are AICIPC. I really hope I can find a source for those cables - it makes identifying problem drives so much easier.

just finished "reuse" Motherboard tray and back window/plate with norco fan plate
the PSU will be 2U 500W.


Motherboard tray fitter


side padding for Norco fan plate


Norco fan plate installed with side padding


Back View

if you want to make those cables by yourself -> /www.aicipc.ru/data/files/files/83_file.pdf

during my investigation:
the backplane is just a dummy plane as I understand where should supports any SATA 1/2/3,
there are jumpers to let external connector control HD LED.
FAILED LED Pins/connectors are on the backplane board, just need to wired to any card that supports failed LEDs


pin/jumper locations

Thanks for that info, bima!

Is that using the configuration it shipped with (well, other than the upgraded Xeons?)? I just finally got around to plugging mine in this morning, and with the two 60mm fans disconnected, two of the three PSUs removed and the two blowers taken out, I'm still pulling 180 watt at the wall with out any drives connected what-so-ever. With all the fans and power supplies connected, I'm looking at 260 watts - still no drives connected. Wish I had thought ahead and ordered the parts to complete this project... I have a nice five day weekend ahead of me, and it's going to be hotter than hell outside, it it would be the perfect opportunity to do this.

Also, mine shipped with an IPMI card in it - AOC-SIMLC. Anybody else get one of these? I've downloaded the manual, Not exactly sure what I can or can't do with this particular card. Apparently I'd need the + model to do KVM over LAN

Finally, do the fault LEDs work with the system as configured (with the Supermicro HBAs) or not?