Array (11+1 drives) running 4.7 started behaving strangely last night and disappeared from my network, although command line control was fine. Rebooted and worked fine for a few minutes before disappearing again.

System is probably 5 (?) years old and still using the original flash drive I got from Tom when I purchased the license. Any way to really test the flash drive to see if it's starting to die? Because of the license issues I can't just copy the files to another usb stick and see if that clears things up can I - at least not without obtaining a new key - yes?

 

I've attached a syslog just to see if there's anything hiding in there that might clear this up-

syslog.txt

Very unlikely it's the flash drive, as once the system boots it doesn't' depend on the USB drive at all (and if the few writes that are made to the drive should fail, I don't believe it has any impact on the overall functionality of the array).

 

I presume, since you said the command line works okay, that you have a display attached to the UnRAID box => does it show anything unusual?    It should just show the logon prompt after a boot, and then nothing more unless it encounters errors.   

 

Are you just losing access to the Web GUI ... or is the array also non-responsive?

 

i.e. if you open Windows Explorer and go to \\Tower  do you get a response?  (This may take a few seconds)

 

Hi Gary,

Nothing unusual that I can see - goes to command prompt and I never lose control/access via keyboard.

 

When array "disappears", I can't do anything over the network. Tower is no longer accessible via either browser or Windows explorer.

When array "disappears", I can't do anything over the network. Tower is no longer accessible via either browser or Windows explorer.

 

But the monitor still just shows the logon prompt?

 

... and, if I understand your original comment correctly, you do things on the system from the monitor (just press Enter and see the prompt repeat;  logon to root;  etc.)

 

This sounds like you're losing the network connection ... but if so, I'd expect to see an error line on the monitor that noted that (but it may not do that if you're only losing an inbound pair on the cable).    Try using a different Ethernet cable ... or at least unplugging/replugging your current one.  Do this with all involved cables [i.e. UnRAID <-> router and router <-> PC]

 

Login prompt is still up - nothing changed, although now I can't do anything with the keyboard (and the array went away also).

 

Unplugged/plugged in cables and swapped out the one for the server - still nothing.

Login prompt is still up - nothing changed, although now I can't do anything with the keyboard (and the array went away also).

 

Unplugged/plugged in cables and swapped out the one for the server - still nothing.

 

"... I can't do anything with the keyboard ..."  ==>  On the UnRAID box?  If so, then clearly it's hung.    So ... you most likely have had a hardware failure.

 

Time to start a bit of troubleshooting ...

 

First question:  When you CAN see the drives, does UnRAID show them all okay?  If you can't see them anymore to answer this, that's okay -- you would have probably noticed if any were "red balled".

 

So ... I'd do the following:

 

(a)  Boot the UnRAID flash and select MemTest.  Let it run for HOURS  (unless, of course, it immediately shows errors => but in the absence of errors, let it run for at least 6 full passes)

 

(b)  Look VERY carefully at your motherboard (use a flashlight) for any signs of deterioration on the capacitors (leakage or bulging).    If you don't see anything there, then I'd try a new power supply for the system.

 

©  If there are no memory errors, and a new PSU doesn't resolve this, then before doing more tests, as a safety precaution I'd unplug the SATA and power cables from all of your disk drives => not strictly necessary, but ensure no additional tests can accidentally write anything to the drives.    Hopefully you won't need to go beyond either replacing defective memory or putting in a new PSU.

 

By the way, if you DO buy a new power supply, get a high-quality single-rail unit.  For the size of your system, something like this would be excellent: 

http://www.newegg.com/Product/Product.aspx?Item=N82E16817151088

Thanks Gary-

Instead of running the memtest, I just swapped out the memory stick for another. Don't see any obvious signs of cap failure (which doesn't necessarily mean anything) - but I think I'll go with the PS suggestion first if the new stick doesn't clear things up. MUCH easier to swap a PS than a MB, that's for sure!

 

Since my MB is 6 years old now - how difficult is it to change it out for a newer one? Simply pull the 2 SATA cards and reinsert into the new board (same positions) and make note of the 4 SATA connections on the MB? (Obviously take a screenshot of the disks/positions for reference as well)

 

Appreciate the help and we'll keep our fingers crossed for now

Swapping out to a new motherboard is very simple.  UnRAID has NO motherboard dependencies, so you simply swap in a new motherboard/CPU/memory combo; connect all the drives; and boot the system.

 

HOWEVER ... with v4.7 the drives have to be in the same relative slots .. and the enumeration of the drives with a new board may be different, so you need to make a note of each drive's serial number, so you can assign them correctly on the new system.  You can avoid this requirement by FIRST upgrading to v5 BEFORE you upgrade the board -- HOWEVER (and this is a BIG HOWEVER) that's fairly risky to do when you're working with hardware that's giving you problems, so I would tend to NOT do that.

 

Hopefully a nice new PSU will resolve everything !!    If it does, you may want to THEN upgrade to v5, so if you do ever move everything to a new board you don't have to worry about which drive goes where