Supermicro : aoc-saslp-mv8 breaks the sleep mode (update - when disks connected)

[redia]

Dear all,

 

I have a problem with a server I am currently building.

nothing really fancy.

running a MSI h55m-ed55 + a few disks..

I had fun with s2ram (using s3.sh) to allow sending my box to sleep...

works like a charm....

Until I installed a brand new supermicro card.. as mentioned in the title it is an AOC-SASLP-MV8 running firmware *.15

but it breaks the wake up.

 

has anyone faced the same issue and/or has an idea ?

 

Cheers,

R

 

 

[redia]

sorry a quick addition to my post.

in fact it works like a charm if the card is plugged....

it is as soon as I connect a drive to it that is stops...

I will do further tests..

 

I hope someone can help because I am lost... lol

 

Cheers,

R

[prostuff1]

Unless someone has the exact same setup as you, everything they are going to say will be shots in the dark.

 

I wish I could help more but, I don't bother with the s2ram, S3 sleep state stuff.

[redia]

Unless someone has the exact same setup as you, everything they are going to say will be shots in the dark.

 

I wish I could help more but, I don't bother with the s2ram, S3 sleep state stuff.

 

Prostuff1,

first of all thanks for the time you took to answer, it is truly appreciated that someone hit the reply button.

 

from my testing I am pretty sure the problem is with the supermicro card. as long as no HD are plugged, it works like a charm..

I am definitely not asking for a straight solution but a few hints but I am sure there are some "pro" around here who might give me a few hints on how to improve the compatibility of s2ram, maybe someone can hint me on the supermicro card bios setting, or even a "generic" MB bios setting that I should make sure I setup.

 

i have tested all the possible configuration of s2ram (p/m/s/ax/v) and nothing seems to work.

so to be honest I do not know in which direction I should look.

 

if someone has tested an alternate controller card which works for s3 and is in the same price range than the supermicro i am interested to..

 

thanks in advance for any charitable soul who can hint me

 

Cheers,

R

[mvdzwaan]

Have you disabled int13 in the Supermicro Bios ?

[eroz]

How do you access bios on the Supermicro aoc-saslp-mv8?

[redia]

Have you disabled int13 in the Supermicro Bios ?

 

mvdzwaan,

yeah I think I disabled the int13.. but I will check and try again in both position if that could change a thing it is worth the try.

I thought it was only to allow bootable device but I might be wrong...

 

 

How do you access bios on the Supermicro aoc-saslp-mv8?

 

eroz

you have to type <ctrl>+<m> right after the disks have been detected.

 

 

R

[mvdzwaan]

As a wake from S3 does share some elements with a boot process, I would always try this. Also as it's an easy check without risk.

[redia]

As a wake from S3 does share some elements with a boot process, I would always try this. Also as it's an easy check without risk.

 

I will take ANY hints

even if it does not work, this is the way to go... someone will eventually be right lol ...

unfortunately it looks like I was right and I already checked that...

I rechecked to be on the safe side.. but no luck..

 

here are the available options in the controller card bios  (in [] are the last tested settings) :

 

Int 13h : enable/[disabled]

silent mode : enable/[disabled]

halt on error : enable/[disabled]

Staggered Spin Up :

          - Number of Devices per group : 1/2/3/4/5/6/7/[8]

          - Group spin up delay : [0]1/2/3/4/5/6/7

HDD-Detect time(s) : 0/1/2/3/4/5/6/7/8/9/[10]

 

 

R

 

 

[spinbot]

I just came across this thread as I wasn't sure if I disabled INT13 (its been a while), so I at least now know how to check.

 

What I also noticed here was the option in the BIOS to stagger the spin-up of the drives.

 

I've never given it too much thought, but does unraid stagger the spin-up off all the drives, when you first turn the server on or would one suggest selecting this option so at least the drives on the controller are staggered?

 

I bought a decent Power Supply ( Corsair 750W single rail ) and it handles the 8 drives I have now, however I am installing 2 more next week, so figured it might be something to ask about 

 

Thanks

 

 

[redia]

I just came across this thread as I wasn't sure if I disabled INT13 (its been a while), so I at least now know how to check.

 

What I also noticed here was the option in the BIOS to stagger the spin-up of the drives.

 

I've never given it too much thought, but does unraid stagger the spin-up off all the drives, when you first turn the server on or would one suggest selecting this option so at least the drives on the controller are staggered?

 

I bought a decent Power Supply ( Corsair 750W single rail ) and it handles the 8 drives I have now, however I am installing 2 more next week, so figured it might be something to ask about 

 

Thanks

 

as far as I understand the stagger spin up from the card is only used during the boot process. this way you limit the amount of power you need during the boot.

as soon as the boot process is over everything depends on the operating system (in this case unraid)

with a 750w psu you should not have any problem with 10 drives.. it depends on which one, but I can hardly imagine that it would be an issue.

I sized my 750 psu for 20+ drives..

 

R

 

[GRipp]

It would appear I have a similar problem.  Getting the server to sleep via the UnMenu 1.3 User Script Sleep button or the custom s3.sh script work just fine at putting the UnRAID system to sleep.  Waking the system up is no problem either, but every time when it wakes up, it fires up all the fans, lights and the four hard drives attached to the motherboard.  However, the lights on the aoc-saslp-mv8 do not light up and the hard drives do not spin up.  I also loose my video monitor, but could care the least about that.  Yes, I have disabled the INT 13h on the cards BIOS.

 

I've been adjusting one setting at a time in both the mother board (ASUS P5G41-M LX2/GB) and the Supermicrop cards BIOS.  Nothing seems to tell the card to kick back on when waking up from S3 sleep.  Is there a command I could add to my sleep script that would tell the card to "reboot" after waking up?  

 

I've got some more research to do, but any assistance would be greatly appreciated.

[GRipp]

Ok.  I messed around for a couple of hours last night trying to get this beast to wake up properly and so far, it's a no go.  Here are a couple of more steps I took to try and figure things out.

 

Changed a lot of Bios settings, when nothing seemed to change the way the machine woke up, I returned it to default settings and made sure the normal settings were in place (WOL,  Boot with USB, etc...).  When that still didn't work, I  checked the Bios version.  It was the oldest version out there.  So flashed the Bios with the newest.  However, no change in the wake up process.  Still just hangs after firing up the hard drives, lights, and fans. 

 

Next step, stopped the array, turned it off and pulled the aoc-saslp-mv8 card all together.  Of course the array would not start as it was now missing three drives.  However, I was still able to put the server to sleep, wake it up, and get back into the web GUI and UnMenu which I was unable to do before.  I was also able to telnet into the system, again, which I was not able to do before.  Before I could allegedly telnet in, and when I did, it would not ask for a password and say "you are connected" but I couldn't do a single thing. 

 

Next I put the aoc-saslp-mv8 card back in but left the breakout cable unplugged from the card, again, the array would not start as it was missing three drives, but I could still put the server to sleep and wake it up just fine.  Plug the breakout cable back in (with the machine turned off) and still the same problem.  Goes to sleep, wakes up, but cannot access the server at all.  This leaves me to believe the breakout cable may be at fault, but I've been using the machine for two weeks or so with zero parity errors (rebuilt parity twice and two more checks) with data on all the drives and would think that if something were wrong with the cable, it would show up before going to sleep and waking up.

 

One update from my previous post, when the machine first wakes up the green LED's on the bottom of the aoc-saslp-mv8 do light up for a second like they do when you first fire the machine up.  But that is it.

 

If no one has any suggestions, I am open to the idea of a new motherboard.  I'd like to use my processor (LGA775 socket Intel 2.3GHz Quad Core) Ram (OZC 8500 DDR2) and of course the aoc-saslp-mv8.  It also has to sleep and wake up properly. 

 

Thoughts?  Ideas?  Wish I could post a system log of what happens when the machine wakes up, but it doesn't even create one until after the fresh reboot.

[BRiT]

I do not know of anyone using the SUpermicro AOC-SASLP-MV8 card that has a working sleep/resume system. It seems that's just the way things are with that card.

[redia]

GRipp,

 

you are at the exact same point as I am

can't seem to figure it out

 

cheers,

R

[GRipp]

I do not know of anyone using the SUpermicro AOC-SASLP-MV8 card that has a working sleep/resume system. It seems that's just the way things are with that card.

 

Really?  No one has this card installed on any motherboard with a proper sleep/resume function working?  That should be a sub note on the Wiki Compatibility Page.  If this can be confirmed, I'll update it myself.  That was really one of the great functions I wanted. 

 

If this is truly the case, I may consider building a second tower for that function to work.  Keep all hdd's on the motherboard and never use a SAS controller card.  Do any controller cards work properly for the sleep/resume function?

[BRiT]

Sleep is nice and all, but when you're using a modern low power motherboard and CPU it's not so important when you have power management setup.

 

I'm talking about an Intel i3 530 or the equivalent AMD setup. When the drives are spun down, you're looking at around 35 watts base for the system, with maybe 8 watts for the additional controller card, then add another 0.8 to 2 watts per drive (depends on make and model). When all 21 drives are spun down on a fully loaded array, you're looking at a range from 64 to 85 watts depending on green or non-green drives.

 

Fans and cooling might add some more watts, but this should give you some ballpark envelope calculations. Modern day motherboards and cpus are amazingly power efficient as are certain hdd's when spun down.

 

Combine this with the headaches of getting sleep/resume properly working, it's obviously hasn't been worth the hassle and frustration for most to work out.

[prostuff1]

Combine this with the headaches of getting sleep/resume properly working, it's obviously hasn't been worth the hassle and frustration for most to work out.

 

This is exactly why I have not attempted to get sleep working on my machine.  It is just not worth the trouble for the very small amount of cash it might save me over the course of a year.

[GRipp]

Combine this with the headaches of getting sleep/resume properly working, it's obviously hasn't been worth the hassle and frustration for most to work out.

 

This is exactly why I have not attempted to get sleep working on my machine.  It is just not worth the trouble for the very small amount of cash it might save me over the course of a year.

 

Honestly, getting the s3.sh script to work was a breeze.  So easy that my wife actually used the WOL functions.  I put a widget on her OS X system, and all she had to do was hit f12 and push the sunshine button.  Fifteen minutes after all the hard disk drives had spun down and the tower pinged our HTPC's and detected nothing would go into sleep mode. 

 

I guess call me old fashioned as I had the "turn the light off when you leave the room" drilled into my head.  Now I know my system will pull 35 - 40 watts at idle which is very little in the big scheme of things. 

 

For now I'll let the tower run 24/7 (as it is also using Transmission and LogMeIn) and see how things go.  If anyone is able to get an unRAID system to sleep and wake properly with the mv8 card properly, let us know.

 

This is a great community for a great piece of software.  Thanks.

[BRiT]

If you read the other S3 / Sleep related threads you'll find out all the frustrations others have had.

 

Even a non-compliant usb device such as a simple freaking mouse can prevent a system from going to sleep. Lately it hasn't been as much of an issue, but if a system is built using hardware older than a year or so, chances are exponentially higher that it will have either a bugged motherboard BIOS or some other drive controller or a NIC or even the video card that won't play nice with sleep and resume.

[skank]

anyone figured this out? i have the same issue

[seim]

same issue, no solution so far