Jump to content

unRAID Server shutting down immediately on power up


davidskirk

Recommended Posts

I have an early unRAID system (about 2 years old) with 11 disks in total varying from 200GB to 750GB (all PATA), running version 3.0 of unRAID.  The system has been reliably supporting me during that time and I love the convenience and flexibility - but now things have changed.

 

I was upgrading 2 disks over the last couple of days to 750GB to get some more capacity.  The first replacement worked fine and the disk was reconstructed.  After making sure parity was fine and all drives were running properly and file system was accessible, I powered down to do the second drive.

 

When I powered up - things got a bit strange.  I normally power up by turning on the bottom power supply (powers the lower 6 drives only) and then switch on the upper power supply (powers the top 6 drives and the MoBo), so that all the disks are powered on when the system is booting.  Well - as I turned on the upper power supply, about 1 second later both power supplies shut down.

 

I tried putting the original disk back in to see if there was something weird happing with the new one, but the same thing happens.  I can now power on only one power supply at a time, but when I switch on the second, both shut down.

 

As each one is working independently (or at least appear to be), I am not sure whether the fault is with one of the power supplies, or whether there is something on the motherboard causing the shutdown.

 

I also am not sure whether the upgrading of the disks actually has anything to do with it - but may be contributing only in the sense that obviously an upgrade process caused me to do more power-cycling than normal.

 

Anyone seen this before - or would know how to determine whether it's MoBo or PS related, and if the latter, how to determine which one might be causing the problem.

 

Obviously I am dead-in-the-water until I get it figured out as the server will just not start at this point.  Any help to avoid me ordering lots of new hardware would be appreciated.

Link to comment

I'm wondering if it is possible that you did something accidentally while you were exchanging the drives.

 

For example, I think that disconnecting the CPU fan will cause a very quick CPU temperature spike and immediately shutdown like you are describing.

 

I'd recommend double checking to make sure that nothing got unplugged or disconnected.

Link to comment

Thanks for the suggestion.  I just checked inside the server box although I am now more confused than ever!

 

If I just turn on one of the power supplies, it powers up fine and I can see the disk slot lights come on and hear the disks spin up.  When I turn on the second power supply, both shut down in less than a second.

 

With the top power supply on, I can verify that the CPU fan, case fans and disk fans are all working - nothing obvious is physically wrong.  Only when I switch on the second power supply does the whole thing shut down.

 

Likewise, if I switch on the lower power supply by itself, it appears to be running fine, powering up all of the lower hard drives.  When I turn on the upper power supply, I then get the same shutdown within one second.

 

The motherboard is therefore not shutting down the upper power supply when that is the only one running.  Strangely, once they have both shut down, if I switch off one of the power supplies (either one), the other one then springs into life.

 

Have never had a machine with 2 power supplies and not sure how the cut-out sensors work, but this is a little strange to say the least.

Link to comment

I am having a similar problem. i have 12 hds ranging from 320 to 1 td (mostly 1 tb and 500 gigs), all seagate baracudas. The power supply is 1000 watt. when i put the 10th and 11th hd in my server started shutting down a second or two after turning it on. at first i thought it was just two much of a strain for  the ps so i unplugged the two hds and started the server. it started fine so i plugged in the two hds and all is well. now i just do that every time i start it. any suggestions?

Link to comment

I am having a similar problem. i have 12 hds ranging from 320 to 1 td (mostly 1 tb and 500 gigs), all seagate baracudas. The power supply is 1000 watt. when i put the 10th and 11th hd in my server started shutting down a second or two after turning it on. at first i thought it was just two much of a strain for  the ps so i unplugged the two hds and started the server. it started fine so i plugged in the two hds and all is well. now i just do that every time i start it. any suggestions?

 

Sounds like you need a better PSU.

 

 

Bill

Link to comment

Well, I have been trying to eliminate the possible source of the problem.

 

1. I have swapped each power supply out (one at a time) - no change - still shutting down after about 1 second.

2. I have disconnected all drives, removed the flash drive - no change.

3. I have disconnected and reconnected all cables on the motherboard - no change.

4. I have disconnected both Promise PCI IDE controller cards and powered up without them - no change.

 

I now have to believe there is something wrong with the motherboard or the CPU - causing everything to shut down.

 

Can anyone help with what motherboard and CPU I should buy to replace them - at this point, I just think it would be easier to do that.  All I can read on the motherboard is Intel and E210882 - but when I search for that online, there appear to be all sorts of boards from all sorts of manufacturers that show up.

 

Does anyone know what board was originally put in there and what CPU - it worked fine for a couple of years, so happy to replace with what was in there originally.  I don't want any compatibility issues and desperately don't want to lose my data sitting on all those drives.  I am assuming I will be able to fire things up with my existing flash drive once I replace these.

 

Any information appreciated.

Link to comment

Have you tried to disconnect the "power button"  (if you have one)

 

If it was defective and stuck closed, I can see it acting a bit like you described.  It is normally a momentary contact switch, so it being held closed would simulate you asking for the power to be turned off forcefully.

 

This is probably not it, since you said you could power up with one power supply, but I figured I'd give you something else to try.

 

Joe L.

Link to comment
I don't want any compatibility issues and desperately don't want to lose my data sitting on all those drives.  I am assuming I will be able to fire things up with my existing flash drive once I replace these.

 

I can reassure you as to the safety of your data.  unRAID is *almost* completely hardware independent, because it uses a well-tested Linux kernel, with good general compatibility, and installs itself fresh with every boot.  That is, it identifies and configures itself for the motherboard, CPU, and all components every single time, as if it had never seen them before.  You can remove all of your drives plus the flash drive, carry them across the country, and re-install in someone else's computer, and the chances are very high that everything will boot and look identical.  It is possible that you *may* need to correct a few device assignments, and it *may* want to run a parity check (which you can stop), but that is all.

 

In the future, I expect to see more users taking advantage of this portability, by installing drives in external 5 and 10 drive boxes, with 1 or 2 port multiplier connections, and traveling with the external drive container(s) and the bootable unRAID flash drive.  This would be an unRAID-powered 10 terabyte portable NAS!

 

I'm hoping Tom will in the future remove the device slot assignments from the disk.cfg file, and just loop through the drives matching their disk numbers with their serial numbers.  That would remove the last hardware dependency I think, and would also be slot independent too, allowing you to renumber your drives at will.

Link to comment

I have the MD-1200 from Lime Tech, so no power button.  Both drives have power switches and that is how I am powering them on and off.  When the system was running, I was of course doing a shut-down from the web admin page.

 

I have now removed the motherboard and tried powering it out of the case with just the main power cable and 12v (yellow and black 4-cable connector) plugged-in.  I still get the same issue (will start and then immediately shut down), so don't think it can be anything shorting out the motherboard.  There really isn't much left to try, although when I removed the CPU heatsink and fan, there was very little in the way of thermal grease (if that's the term!) on it.  Is it possible the heatsink could not be conducting heat properly and causing CPU to overheat, shutting everything down?

 

Can anyone suggest a motherboard/CPU combination that can drive 10 PATA drives with the Promise controllers?  Just want to get this thing back up and running - every day is making me more nervous - but do realize that my data is almost certainly still in good shape on the disks at this point.

Link to comment

I have the MD-1200 from Lime Tech, so no power button.  Both drives have power switches and that is how I am powering them on and off.  When the system was running, I was of course doing a shut-down from the web admin page.

 

I have now removed the motherboard and tried powering it out of the case with just the main power cable and 12v (yellow and black 4-cable connector) plugged-in.  I still get the same issue (will start and then immediately shut down), so don't think it can be anything shorting out the motherboard.  There really isn't much left to try, although when I removed the CPU heatsink and fan, there was very little in the way of thermal grease (if that's the term!) on it.  Is it possible the heatsink could not be conducting heat properly and causing CPU to overheat, shutting everything down?

 

Can anyone suggest a motherboard/CPU combination that can drive 10 PATA drives with the Promise controllers?  Just want to get this thing back up and running - every day is making me more nervous - but do realize that my data is almost certainly still in good shape on the disks at this point.

I have the exact same configuration.  An original MD-1200 with two power supplies, and no power button, as assembled by Tom.

 

Your data is fine...

 

It is possible the CPU would overheat, but very unlikely in just a few seconds.  There should be very little heat-sink compound... Most "intel" guides describe using the amount the size of a grain of rice, spread smoothly across the heatsink surface contacting the CPU.  It is there to fill in the voids in the metal, not to fill gaps.

 

Joe L.

Link to comment
There really isn't much left to try, although when I removed the CPU heatsink and fan, there was very little in the way of thermal grease (if that's the term!) on it.  Is it possible the heatsink could not be conducting heat properly and causing CPU to overheat, shutting everything down?

 

Lack of thermal grease, or a 'cracked' thermal layer, can shut your system down quickly.  If the thermal grease 'layer' or pad is rather old, and you jarred the CPU heatsink, you may have broken the seal or integrity of the layer, and it won't conduct heat fast enough away from the CPU.  But I thought you said above that it was running fine at times, at least with one power supply.  An ineffective thermal grease or pad may not run long under any conditions.

 

You might try cleaning it up, and applying Arctic Silver or comparable.  See this post through this post, and the ones in between, for tips on using Arctic Silver.  The cleaning instructions are VERY important.

Link to comment

Archived

This topic is now archived and is closed to further replies.

×
×
  • Create New...