See my signature for build info.

 

Booting my box has been very flaky.  It was working mostly the other day, though I had some issues seeing some files with XBMC (diff issue, see below).  So I rebooted it.  It didn't boot - it was just sitting at the first BIOS screen, doing nothing.  I rebooted again.  Same thing.  I pulled the unraid flash disk and rebooted.  It progressed as normal, getting past the first BIOS screen, but eventually stopped of course because it couldn't find an OS to boot.  I powered down, put the flash back in, and powered up again.  Now it booted fine to unraid OS and I could access the web UI.

 

For grins, I cycled it again.  It appeared to boot okay, got me to a login prompt as usual, but now I couldn't get to the web UI nor could I telnet in - but it was on the network and was assigned the proper IP (reserved), according to my router.  Not sure what that was about.

 

I cycled it again.... it got stuck at the BIOS again.  Without rebooting, I pulled the flash drive... it began to proceed again.  I put the flash back in immediately, and it booted as usual and I could get to the web UI and telnet.

 

I've actually got two other problems, but they may be unrelated.  I'll post them separately:

- unmenu won't start, but prints no errors

- file permissions are not working for XBMC

 

 

(Again, see my signature below for setup details.)

 

I did some more playing around with my unraid box tonight, doing lots of reboots and trying to collect data.  It's hard because when it fails to boot, or partially boots, you can't save the syslog anywhere for later analysis. I was able to save one (below), but not sure if it will help.

 

There's definitely something wrong with either the USB bus/chip, the USB driver, or the USB flash drive.  Very often my boot would hang at the BIOS screen, and I couldn't get any keys to work (so, either a USB problem or the BIOS just never got far enough to enable USB).  And to fix it, all I had to do was pull the flash drive out.  In most cases, it would continue to boot.  I could re-insert the flash drive immediately and usually it would boot.  (There's a couple intermediate bios screens where it shows you the drives and something, then you get to the boot screen where it says it will boot unraid in 4 seconds, or I could choose memtest.)  On at least one occasion, when I pulled the flash drive the whole PC shut down immediately (like I pulled the power cord) - it was stuck at the BIOS screen till I pulled it.

 

I haven't tried all the USB ports yet - that could be a solution.  There are a bunch on the back and a couple on top of the box.  I've tried a couple, though, and it doesn't seem to matter.

 

I think I managed to get 2, maybe 3, reboots in a row that came all the way up.  Many got hung at the BIOS screen.  Some booted the image, the logs started rolling, and then I saw one log about not finding the UNRAID drive... though after a long pause, I got a prompt and could log in (but the box was unusable).  But I couldn't access /boot or /mnt/diskX or anything, so I couldn't save the syslog anywhere.  I couldn't FTP or telnet in, either.  In a couple cases, it seemed to boot all the way up, beeped twice (which is the last thing is usually does on a clean boot), but I couldn't access the web UI.  I think I managed to save that log, and I have yet to debug it.

 

And I'm actually worried that there are even issues once the box boots completely and seems to be running.  If the flash drive becomes inaccessible... well, what's on there?  The OS would be in RAM at that point.  I guess the array config lives there, so I wouldn't be able to tweak anything... but normally I would shut down the box first, anyway.  I wonder if it stores anything about the parity check on the flash drive. 

 

(There really needs to be some extra ways to save or transfer the syslog on failures - how else can you debug?)

 

So, my question really is this: how do I debug this issue to figure out what's really going on here?

I used to have intermittent problems while booting (i.e. no NIC found). Turns out it was the USB drive itself. You can troubleshoot it by trying different USB ports, but I think the drive is the likely culprit. Note that Windows didn't show any obvious problem with the USB drive.

Since I'm still under the "free" license limit, I should try another thumb drive.  What do I need to copy over?  The /config directory?

 

However, I have already paid for a license, in anticipation of adding some more drives that I have laying around.  Is there any way to transfer a license?  I'm guessing not, but I'm kinda screwed if the drive is bad.

Since I'm still under the "free" license limit, I should try another thumb drive.  What do I need to copy over?  The /config directory?

 

However, I have already paid for a license, in anticipation of adding some more drives that I have laying around.  Is there any way to transfer a license?  I'm guessing not, but I'm kinda screwed if the drive is bad.

 

When I realized it was my USB drive that was the issue, I emailed Limetech support and they gave me a new key for a new USB device. They were great.

I'm guessing it's the USB drive.  I'm going to try a different one tonight. 

 

Here's a snippet of logs that are all over my syslog...

 

May 26 12:09:34 Tower unmenu[1347]: df: `/boot': Input/output error
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device
May 26 12:09:34 Tower last message repeated 2 times
May 26 12:09:34 Tower kernel: FAT: FAT read failed (blocknr 2082)
May 26 12:09:34 Tower unmenu[1347]: df: `/boot': Input/output error
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device
May 26 12:09:34 Tower last message repeated 2 times
May 26 12:09:34 Tower kernel: FAT: FAT read failed (blocknr 2082)
May 26 12:09:34 Tower unmenu[1347]: df: `/boot': Input/output error
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device
May 26 12:09:34 Tower last message repeated 2 times
May 26 12:09:34 Tower kernel: FAT: FAT read failed (blocknr 2082)
May 26 12:09:34 Tower unmenu[1347]: df: `/boot': Input/output error
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device
May 26 12:09:34 Tower last message repeated 2 times
May 26 12:09:34 Tower kernel: FAT: FAT read failed (blocknr 2082)
May 26 12:09:34 Tower unmenu[1347]: df: `/boot': Input/output error
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device
May 26 12:09:34 Tower last message repeated 3 times
May 26 12:09:34 Tower kernel: FAT: FAT read failed (blocknr 2082)
May 26 12:09:34 Tower kernel: sd 4:0:0:0: rejecting I/O to offline device

I just commissioned a new USB stick for my unraid box, and I was able to reboot from the web UI 5 times in a row with no problems.  I've never done more than 2 in a row on the other USB stick.  So, I'm 95% sure at this point that the problems I was having pertained to the USB drive.

 

The drive I'm using now was a give-away drive from my company... not exactly name brand (unraid UI says it's "CBM").  But it seems to work.  I'm wondering if I shouldn't still buy another thumb drive, something more reputable, before trying to convince Lime to give me a new license.  Any thoughts on this?

Ultimately up to you.  I continue to use a 1 GB drive that I claimed from a lost and found where I used to work.  No name brand at all, unRAID UI even calls it 'Generic' 

 

If I were buying a new flash drive I would definitely get something reputable, but as long as it seems pretty reliable it shouldn't matter.  5 reboots would satisfy me.  If/when it dies you can always contact LimeTech to transfer your license, but there may be a bit of server downtime involved.