I'm in the middle of a server upgrade and I have run into a multitude of problems. I've been slowly knocking them out one by one as I go along but I'm at my wit's end with this last problem.

 

Here is what I added: I will put my complete hardware list at the end if that will help

1. PSU - SILVERSTONE Strider series ST60F-PS 600W ATX12V

2. 4 new 3TB drives WD30EZRX (one as a new parity)

3. Bitspower X-Station Power Extension II (2 into 8 4-pin molex connector block)

 

After extensive testing and retesting I have isolated the problem but I don't know how to fix it. If I power up the machine with everything connected (including a monitor to see what's going on) it will go straight to the motherboard splash screen and stay there indefinitely. If I enter Bios I find that my USB boot drive is missing, therefore it will not boot.

 

If I disconnect the power to all drives and restart, then it will boot into unRAID without any problem every time. I then reconnect power to the drives and it will recognize (most) of the drives. Every time I try this there are different drives missing, one time it will be the parity + 2 others, sometimes it's only 1 drive. Only the drives connected to my Supermicro (AOC-SASLP-MV8) card are affected. The 6 drives connected directly to my motherboard have no issues. At one point about a week ago I had the machine up and running. I was using 2 PSU's, one for the MB and one for the drives. I was able to install the new parity drive and rebuild a failed drive from the new parity. I started moving files around to different locations and everything was going smoothly for several days. I figured I would get all my file shuffling out of the way and then address the previous issue when I was done. I had an issue with some files not showing up in my share so I restarted and now I have the aforementioned problem with the drives missing. When I click on the drop down box that says "No Device" in the web GUI there are no drives to choose from that I could add, it's as if they are missing completely.

 

I have already replaced my motherboard early on in this process thinking that the missing boot drive was an issue with that but the new one did exactly the same thing ($70 wasted). I have a feeling that my Supermicro card is the culprit in this scenario but before I throw another $100 at this thing I would love to hear your thoughts and/or suggestions.

 

If I left out any pertinent information please ask and I will do my best to answer. I really need to get this thing back online it's been about 2 months so far.

 

SYSTEM:

OS: Server Pro v.5.0-Beta 14

MB: Asrock 880GM-LE FX

CPU: AMD Sempron 145 Sargas

PSU: SILVERSTONE Strider ST60F-PS 600W

FAN CONTROL: Lamptron FC Touch

HOT SWAP CAGES: Icy Dock 4 in 3, & 3 in 2

RAID CARD: Supermicro AOC-SASLP-MV8

STORAGE: 11 X WD Green drives 2 & 3TB mixed

CACHE DRIVE: WD 500GB Caviar Black 7200RPM

I was using 2 PSU's, one for the MB and one for the drives.

Possibly too weak PSU?

 

Try this:

Let the rig power up and when all drives are fully spinning press the "reset" and see if it boots up normally.

Or, reduce the drive count for testing.

Or simply try another PSU.

That's how I got it up and running a week ago. I started the drives with a 400w PSU and then booted the unRAID with the new 600w PSU and it started right up. Since I shut it down a while back I have not been able to get it back online. I keep getting missing devices in different slots. I will try your suggestion about reducing the # of drives. I believed that 600w @ 49A on a single rail would be enough to run 14 total drives.

 

This thing is driving me insane. I don't mind replacing a part if it is needed but I can't afford to replace every part just to see if that will work.

I know very well, how annoying these things can be.

Although your PSU is new and should have enough steam for your configuration, you

can't rule out a defect or simply incompatibility!

Of course you cannot rule out the HBA to be the problem as well, but if you manage to get it running

with 2 PSU's it's most likely that there is the key.

Can you reliably start with 2 PSU's?

 

PSU's can cause nasty problems.

A PC that I built for my brother long time ago suddenly wouldn't boot anymore.

He had to fiddle around and randomly it came up or not.

I swapped the mainboard just to experience exactly the same behaviour.

After talking to a friend of mine who's also kind of a problem solver in his circle of friends

advised me to swap out the PSU. Problem solved.

My brother in law, who's a electronic technician also told me about PSU's that produce harmonic waves that can interfere with motherboard components and cause instabilities.

They did some testing at the institute he was working.

 

After extensive testing and retesting I have isolated the problem but I don't know how to fix it. If I power up the machine with everything connected (including a monitor to see what's going on) it will go straight to the motherboard splash screen and stay there indefinitely. If I enter Bios I find that my USB boot drive is missing, therefore it will not boot.

 

 

 

 

When you enter the BIOS, do you see all of the other drives listed (that are attached to the Supermicro Card) in the boot options?

 

The BIOS might be running out of boot options and with that many drives listed, it is dropping the USB from the boot order.  Try going into the supermicro Bios (Ctrl-M?) and disabling Int 13h which will disable the supermicro card from being able to boot.

 

After you get the system to boot reliably with all of the drives connected, then post a syslog if any drives are missing / disabled.

 

As an aside, rather than using your Bitspower X-Station Power Extension II to split 2 molex power connectors to 8, you would be better off order extra modular cables from silverstone to better distribute the power from the supply.  All splitters will introduce voltage drops, and you may also be exceding the capacity of the connector feeding it.  (When I bought my Antec modular supply, there wasn't enough molex connectors, so I just emailed the company, and they sent me a couple extra cables at no charge)

Can you reliably start with 2 PSU's?

 

No, I put my original 450w PSU back in and had the same problem. That's 3 different PSU's now, a 400w, 450w, and the new 600w.

 

 

When you enter the BIOS, do you see all of the other drives listed (that are attached to the Supermicro Card) in the boot options?

 

No, none of them are there.

 

The BIOS might be running out of boot options and with that many drives listed, it is dropping the USB from the boot order.  Try going into the supermicro Bios (Ctrl-M?) and disabling Int 13h which will disable the supermicro card from being able to boot.

 

I tried this and it finally worked. I am able to boot to unRAID and add the missing disk but now I have some other problems. I will make another post with pics after I address your points in this one.

 

After you get the system to boot reliably with all of the drives connected, then post a syslog if any drives are missing / disabled.

 

I don't know how to do that, please advise.

 

As an aside, rather than using your Bitspower X-Station Power Extension II to split 2 molex power connectors to 8, you would be better off order extra modular cables from silverstone to better distribute the power from the supply.  All splitters will introduce voltage drops, and you may also be exceding the capacity of the connector feeding it.  (When I bought my Antec modular supply, there wasn't enough molex connectors, so I just emailed the company, and they sent me a couple extra cables at no charge)

 

I had already made a custom jumper a few days ago with 16ga. wire (bypassing the power block) and it made no difference.

 

After you get the system to boot reliably with all of the drives connected, then post a syslog if any drives are missing / disabled.

 

I don't know how to do that, please advise.

 

 

http://lime-technology.com/wiki/index.php?title=Troubleshooting#Capturing_your_syslog

 

If would probably be good to also post a screen shot of the drive assignments showing the missing / redballed disks

I had already made a custom jumper a few days ago with 16ga. wire (bypassing the power block) and it made no difference.

What do you mean by that?  Did you make your own 2 to 8 molex splitter?  If so, then for all intents and purposes it is the same thing, and will have the same issues if this is a power issue caused by too many devices on the wires (regardless of whether or not the supply is single rail)

I have it up and running but one disk (#4) is in the process of rebuilding and another disk (#5) is completely missing even though it shows up as green

 

My supermicro card has Led's lit up but I don't understand what they can mean because the lights dont correspond with the failed or missing disks.

 

The Parity through Disk 7 are all the drives on the Supermicro card. 8 through 12 + cache are all motherboard SATA drives. 8-9 are empty slots at the moment, I am saving those for future additions.

 

Supermicro LED's

I think I finally have it straightened out. The problem turned out to be something I never suspected, my Icy Dock drive cage.

 

I pulled disk 5 out of the slot and put it in another one just to see if it would power up and it did. Then I took it back out and laid it on the desk. I put the SATA cable in it and ran a new power cable then it booted right up with all disks present and no errors.

 

The device light in the web GUI is still showing red but all the files are there and available. I'm going to leave it running for a while while I go to see that John Wick movie and then try a power down and reboot.

 

The drive cage is about the same price as the Supermicro card but I really don't mind spending the $$ now that I know what the problem is.

 

Fireball & Squid thank you so much for your help.

 

The device light in the web GUI is still showing red but all the files are there and available.

That means the drive had a write to it fail, so it was taken offline and the array is now running degraded, that drive is being emulated by parity + all the rest of the drives.

 

http://lime-technology.com/wiki/index.php/Troubleshooting#What_do_I_do_if_I_get_a_red_ball_next_to_a_hard_disk.3F