[Edit: I'm an idiot: my box has two HDDs, one data, one parity, not three as I originally wrote!]

 

Hi, my unRAID box stopped working properly yesterday and I'm lost as to how to diagnose the problem.  My box is running unRAID 6.2.4.  Installed are three two 4 TB discs (one parity, two data), a 60 GB SDD cache drive, and the obligatory flash drive.  This is all on an Intel G1840 with 4 GB RAM.

 

The symptoms:

- the box will not boot properly;

- I can ssh into it and run bash etc. (although PuTTY warns that the ssh keys have changed, which it's never done before, and I'm not asked for the password, which I always had to provide in the past);

- however there is no web UI;

- running `ps aux | grep -i emhttp` comes up blank;

- there is nothing obviously alarming in /var/log/syslog;

- /boot/config/plugins/ contains only /dynamix/ and that is empty -- which is odd, because I definitely had the Dynamix 'sleep' plugin installed and running;

- after a few hours trawling the forums for advice, I tried this magic incantation in bash: `nohup /usr/local/sbin/emhttp &` which did succeed in getting the web UI up and running;

- however, I was required to re-enter my licence key URL (the GUID for my USB flash drive was recognised correctly);

- all my configuration is missing -- it's as though I have a blank system again;

- I can re-start the array with the two data HDDs, but it can't find the parity HDD (i.e., nothing is listed in the drop-down menues on the 'Main' web tab);

- if I reboot after doing this, it still won't start the web UI;

- all my Docker apps (Plex etc.) appear to be gone -- at least, there is no Docker tab on the web UI.

I've gone to the web UI Tools > Hardware Profile page and clicked 'Upload'.

I've gone to the web UI Tools > Diagnostics page, clicked 'Download', and attached the resulting ZIP file to this message.

 

I'd be tremendously grateful for any help at all at this point!

Ralph

tower-diagnostics-20170321-1416.zip

Edited March 22, 20179 yr by ralphbecket
Stupid mistake...

Have you tried putting your Flash Drive into another machine and attempted to boot it? I wonder if your Flash got corrupted? 

Sorry I haven't had time to read your Diagnostics. Just a quick stab in the dark. 

@kizer I haven't tried booting another machine with the flash drive, but I've plugged it into my windows box to take a backup -- everything appeared fine there.

Definitely a problem with your flash. It has a lot of files from "repairs" (*.REC) and a lot of other files missing. Do you have a backup?

47 minutes ago, trurl said:

Definitely a problem with your flash. It has a lot of files from "repairs" (*.REC) and a lot of other files missing. Do you have a backup?

 

If you have a backup - be careful about restoring it. If you have made configuration changes to your array since the backup was taken, the backup could put your disks back into the old configuration. So if, for example, you upsized your parity and assigned the parity disk as a data disk, restoring the backup would make the data disk the parity disk again, and corrupt that disk when the array is started (which often happens automatically)!

 

There is a tiny little file called super.dat that you might try to rescue from the corrupt disk USB as it stores all of that type of information.

Many thanks for this!  So it's definitely a corrupt flash drive?  I wonder how that could have happened.

 

I've found a super.dat in the backup I took of the flash drive the other night: %unRAID Backup 2017-03-21%\logs\tower-diagnostics-20170205-1532.zip\tower-diagnostics-20170205-1532\config\super.dat which is a 4 Kbyte binary file of some kind with a modification date of 2017-02-05 -- I haven't changed anything other than the sleep plugin settings in over a year and a half.

 

Assuming I can use this file to recover my previous state, can anyone tell me in simple terms what I should do next?

Was the backup taken with the array stopped?

 

super.dat only contains your drive assignments, and whether the array was stopped or started. When it boots, it will check the start/stop recorded in the file and if it sees the array was started, it will assume an unclean shutdown (since super.dat is saved with the array stopped when shutting down cleanly) and begin a parity check.

 

You can prepare the flash like a new install and copy the config folder from your backup.

I took the backup by physically removing the USB flash drive when the unRAID box started acting up (i.e., had lost the array, wouldn't boot to the web UI, etc.)

 

I will try restoring the backup copy of super.dat to the flash drive config directory and see if that works.

 

Thanks again for your help on this one.

7 hours ago, ralphbecket said:

I took the backup by physically removing the USB flash drive when the unRAID box started acting up (i.e., had lost the array, wouldn't boot to the web UI, etc.)

 

I will try restoring the backup copy of super.dat to the flash drive config directory and see if that works.

 

Thanks again for your help on this one.

I'm not sure if you fully understand. I was suggesting you do a new install and copy your whole config from backup, not just super.dat. The part about super.dat was just a warning about some issues people have from restoring that specific file.

Sorry for the delay in responding (work, life, etc.).  I copied over a backup on to the USB stick and, thank heavens, we're back in action -- many thanks indeed!  I have a suspicion a series of power cuts may have been responsible, but either way, this is a good result and everybody's help is much appreciated!