3 passes in memtest and no errors.  How long should I let this run?

I'd let it run overnight.  Your machine locking up/crashing takes hours.

Joe - I didn't set anything special.  Just left things set to auto.  I could try the other Corsair RAM I have and see what happens. 

I looked at your MB manual, unless I missed it, it did not have any manual settings for voltage, timing/clock speed... It just had specific RAM listed that would work.

 

Yes, try the other RAM.  You know what you are currently using is not what it specified.

Ok, I'm going to let it run overnight and see what happens.  I may try the other RAM even if it doesn't fail.  Stay tuned.  Thanks.

I ran memtest overnight and got 8 passes with no errors.  The other RAM I have is 4GB of Corsair XMS2-8500 1066mhz.  Maybe tonight I will switch the RAM and check the BIOS settings one more time for any clues to the unresponsive issue.

I ran memtest overnight and got 8 passes with no errors.

That is encouraging.

  The other RAM I have is 4GB of Corsair XMS2-8500 1066mhz.  Maybe tonight I will switch the RAM and check the BIOS settings one more time for any clues to the unresponsive issue.

It can't hurt to try.

 

Joe L.

I guess I'll swap the RAM tonight.  Any other things to check?

I plan to look through the BIOS again.  Hope that the RAM swap stops the sleep or crash issue whatever it is.

 

In the BIOS, make sure you do not have any power-off times scheduled.

With regards to 'shut down' or 'sleep', normally if a server 'sleeps' it would resume for the state in which it was last working in, do you know for sure if this is the case? it seems like it is powering on from the start not waking up from a sleeping state (just guessing). It could be it is crashing due to some hardware fault, which could therefore be RAM corruption or in other cases the system board or CPU. We had one member who went though hoops trying to resolve his crashing issue, only to find that the mobo has the problem.

 

I guess I'll swap the RAM tonight.  Any other things to check?

I plan to look through the BIOS again.  Hope that the RAM swap stops the sleep or crash issue whatever it is.

 

I noticed that my Supermicro motherboard has this feature called watchdog.  It appears to be able to reboot the  system when a software application hangs.  The manual says the jumper needs to be set as well as enabling it in the BIOS.  I'm going to disable the jumper.  It is already disabled in the BIOS.  I also reflashed the motherboard's BIOS just in case something was strange.  I swapped the RAM to the Corsair.  I'm going to run a memtest with the new memory then let it sit overnight and see what happens.  Also disabled ACPI in the BIOS.

The memtest on the new RAM was fine, no errors.  The last parity check on my server reported no errors.  This morning I checked to see if the system had gone to sleep or if it was off and so far so good but it was a shorter period of time since I fell asleep during the memtest last night then woke up and booted to UNRAID.  So while it looks good I'm not prepared to say it's fixed yet.

For a proper test, leave your server on for several days at end, then if you're satisfied that it is sorted, commence another several day running test, consisting of random tasks to it, like reading/streaming from it and writing data to it as well. If after those two lots of tests and the server stays on, I'd say it should be fine. Keep us posted.

The server is always on.  I usually notice the issue when the server sits overnight and I try to access the shares during the next day.  The shares cannot be found and I would have to power off and back on the server.  We'll see what happens over the next few days.  Hopefully it will just stay on now.

The server is always on.  I usually notice the issue when the server sits overnight...

Is your server hooked to a UPS?

 

Right now it's not.  I don't really have power issues where it's plugged in but I do plan on getting a UPS sometime soon.  I actually need a couple of them throughout the house.

I don't really have power issues where it's plugged in

And how do you know that?  Power surges occur all the time. And they can cause the trouble you are describing.

 

Because none of the other equipment I have plugged into that same area have any power issues.

For the interim, i'd invest on buying a power point line filter, these are obviously not going to be as good as a UPS but it'll help in filtering the power flow and can protect your connected equipment from potential power surges and brown outs. For each power point they are mo more than about $10 and it is some assurance i guess. All of my PC's and servers have a surge protector on it.

 

Because none of the other equipment I have plugged into that same area have any power issues.

Well so far so good, the server has stayed available.  Going to give it another try when I get home after work today.  Since changing the RAM, flashing the BIOS and turning off any power management settings it seems to be stable now.  All parity errors seem to have gone away as well.  Hopefully it is fixed now and I can enjoy the huge storage boost I am getting with UNRAID. 

 

I have one more spot for a drive in my server.  I'm not sure if I should put another data drive or a cache drive in.  If I go with a cache drive would I be better off putting a fast drive like a raptor or velociraptor?

If I go with a cache drive would I be better off putting a fast drive like a raptor or velociraptor?

 

Nope... won't help at all unless you are writing files locally on the server itself. 

 

Otherwise, you are limited by the LAN speed, not the speed you can write to the cache disk.  In the same way, when moving files from the cache drive to the protected array you are limited by the write speed of the protected array, not the read speed of the cache drive.

 

 

Well the server is still awake.  Seems to be working as I want it to now.  Thanks for all the help.  I guess the issue is resolved.

Good to hear, buggy hardware can make a unraid rig do weird things, hence why the RAM testing and what no can assist in working out what the actual fault can be. I'd invest in a decent surge protector for your rig, just to be sure to be sure and to have a good flow of electricity.

 

Well the server is still awake.  Seems to be working as I want it to now.  Thanks for all the help.  I guess the issue is resolved.