Anyone with working knowledge of Rackable SE3016 chassis?

T

themaddhatter

n00b
Joined
Jun 2, 2011
Messages
8
Here is the gear:

- (2) rackable chassis (one chassis is in service full of 2 TB; other to be put in service as soon as space demands warrant)
- HP proliant N36L microserver
- LSI SAS-3801E HBA
- WHS 2011
- Drives pooled & software raid using FlexRAID

The problem:

- Server setup is NOT required to be on 24/7/365, so I can hibernate it after X amount of inactivity, then wait a short while for it to come to life.
- Problem is, if the system hibernates, the Rackable keeps rocking along, spinning all 16 drives, wasting $ like gang-busters.
- The rackables SUCK POWER! I mean, a LOT. Power consumption increased 50% at the meter as a result of this chassis alone. HUGE HOG. This equates to ~ $70/month with just the addition of the first chassis.

What I want to do:

- I want to set up the rackables to do the following things:

A: Power down when server goes into hibernation
B: Power up when server comes out of hibernation

Ideally, the SATA drives would just spin down like they used to when I had these drives in eSATA boxes, but I basically ran out of usable eSATA and PCI-X slots (hence the reason for going to SAS). But, LSI doesn't support MAID or drive spindown (directly from their support personnel), so I have to do this a different way.

I wanted to find out if I could configure the Rackable controller to autostart with power on, or what I needed to access any internal functions using the 9 pin RJ jack on the front of the chassis, but the support people at Rackable don't care unless you have a support contract. Well, SUCK IT SGI, AND YOUR INDIAN SUPPORT STAFF WHO DON'T UNDERSTAND ENGLISH!... sorry :eek:

Here are my questions to all of you

1: Has anyone been successful at setting up serial com with their SE3016? If so, what did you do?
2: Is there a way to configure the controller to either spindown drives, or autostart on power?
3: If not, has anyone ever built a logic circuit to send a momentary pulse to engage the power circuit?

If I can get the chassis to autostart on power, I can set up either a USB controlled power plug or use an IP-controlled power strip to enable power after the server wakes from hibernation, but as it currently stands, I would still have to push the "power" button on the rackables to bring them back out.

And no, I don't want to just "short" the two switch wires together, as these are never meant to be constant powered, only momentary, and I want to still be able to restart/power off the chassis using the power button for service/drive swaps.

Any info would be greatly appreciated.

I
 
That is fine & dandy, but it won't turn on the chassis (that's the problem).

The chassis has a momentary on switch, just like a regular computer. In fact, it has an ATX power supply in it.

On a regular computer, the way that autoboot works is it is enabled in the BIOS. On this chassis, I currently have no way to access the controller (or even if it is accessible) to even investigate if it has the ability to autoboot on power.

So, I need to either find a way to get in there & set it (if at all posslble), or ghetto rig a circuit which will send a momentary pulse of X power to the controller to boot up, but not send additional pulses after it sends one. Then, when the power is lost to the device (via IP or USB controlled AC power plug), the unit just shuts down.

See the issue here? :D It would be easy if the PSU was a constant switch-on style (old school), but since it is momentary on (like modern PSU), I have to deal with this momentary on jigaboo (my highly technical term for stupid crap that shouldn't be stupid).
 
http://www.pwrusb.com/powerUSB-digitaI0.html

This is a USB controlled power strip with 5 digital I/O lines. In addition to being used as a computer controller strip, you can use it to do things like trip switches, etc. Plug the usb end into your server, and you can use it to trip the power switch (You will have to run a pair to the enclosure you wish to control and put it there in place of/in parallel with your momentary switch)
 
Last edited:
I have an SE3016 and I'm also in a similar situation.

The SE3016 uses an ATX power supply, and can be powered on and off by grounding the green wire on the ATX connector. A circuit like this (http://www.oldradio.com/current/cont2mom.htm) with 12v being supplied from the server might work for automatically powering on the array but there is a problem. In an ideal world if the servers 12v loses power it means its been shutdown/hibernating but that is not always the case. For example I'm not sure if during a reboot the 12v might be cut momentarily then power back on, re triggering the relay and powering off the array when it should really stay alive. A workaround would be to make the array not be able to power off from a button press once it is powered on. By tapping into the array's 12v you could have a normal close relay cut the connection from the green wire when the array powers up, making it unable to power off by grounding the green wire. The only way to power it off would be to then use a network controlled power switch.
rlya.png


I'm not an electrical engineer or anything so if this is completely ridiculous or won't work give me a bit of a break.
 
Emtu said:
I have an SE3016 and I'm also in a similar situation.

The SE3016 uses an ATX power supply, and can be powered on and off by grounding the green wire on the ATX connector. A circuit like this (http://www.oldradio.com/current/cont2mom.htm) with 12v being supplied from the server might work for automatically powering on the array but there is a problem. In an ideal world if the servers 12v loses power it means its been shutdown/hibernating but that is not always the case. For example I'm not sure if during a reboot the 12v might be cut momentarily then power back on, re triggering the relay and powering off the array when it should really stay alive. A workaround would be to make the array not be able to power off from a button press once it is powered on. By tapping into the array's 12v you could have a normal close relay cut the connection from the green wire when the array powers up, making it unable to power off by grounding the green wire. The only way to power it off would be to then use a network controlled power switch.
rlya.png


I'm not an electrical engineer or anything so if this is completely ridiculous or won't work give me a bit of a break.
Click to expand...

Nice! That is what I am thinking.

Also had someone recommend a 555 monostable circuit. Will investigate & get back to you on what I find.

I also had someone say that "grounding out the green wire on the PSU doesn't damage anything".

Am exploring setting that up as well connected to a latching relay & a momentary switch, so that I can power down with the power button as well as automate the power down.

Once I have that stuff figured out, then it is a USB control circuit (either internally using the USB voltage, or a USB-controlled AC plug).

Shane
 
I don't know if this is still relevant, but getting the chassis to auto-power on after being plugged in is as simple as jumpering the two pins on the little power board. The front power switch still functions normally.
 
Sorry to resurrect an old thread, but this is the only reference i've found to this.
I have just bought one of these chassis and want it to power on after an outage.
The last post mentioned jumpering some pins, which pins exactly? Perhaps a picture?
Thanks
 
I had tried everything with the jumpers, and found it easier to make my own circuit.

After dicking around with discrete circuits (several iterations), I got tired of breadboarding stuff which simply didn't work for me. So I built the controller using an attiny85. All the bits would be less than $10, but most of the stuff was kicking in my shop. Couple hours writing program, couple hours breadboard & solder, and it was done.

-Clipped the 4 wires to the front controller (Switch In & Out, Led + and -) and the four from the PSU (PS_ON, PS_OK, PS_Standby, and Ground)
- Bought a prototype PCB from RadioShack (the two-piece ones that snap in half) which allows it to mount to the original controller space (just used one hole to hold down)
- Programmed an attiny85 (flashed by an arduino) to sink the PS_ON circuit on power-up (hence turning on PSU), and uses the switch to control whether the PS_ON pin sinks or goes "High" (i.e shutting down PSU).
-PS_OK controls the front LED to tell me if the PSU is working fine
- Cut a USB cable in half, stripped & soldered the ends to the front & rear wires, and soldered usb headers to the pcb for connectors.
- Couple resistors & caps, and that was it. Works well, and am happy with the result.
 
I too have tried all combinations i could think of with the jumpers, thats why i had to ask :p

Anyway, thanks for the tip, i'll see if i can do something similar.
 
evolver56k said:
I don't know if this is still relevant, but getting the chassis to auto-power on after being plugged in is as simple as jumpering the two pins on the little power board. The front power switch still functions normally.
Click to expand...

I would be interested in knowing what 2 pins you jumped as well. I think I've tried just about every combo, and nada.
 
I may be incorrect, but i believe there is more than one revision of this enclosure in the wild. While the expander board appears to be similar enough, the other control circuitry may differ.

EDIT: turns out i probably AM wrong :p i've been meaning to tear this critter apart and mess around with it but haven't had the time or inclination to touch it, plus it's so neat and pretty inside...
 
Last edited:
You must log in or register to reply here.
Back
Top