Flash drive died - Unraid still up

[jasonfox]

I'm in almost the same situation as teedge77 in this thread.  http://lime-technology.com/forum/index.php?topic=54189.0

 

I've already secured a new flash drive, installed a fresh copy of Unraid and the pro.key that was assigned to dead flash drive on it. I just want to be sure I'm following the correct procedure going forward.

 

Even though Unraid is still up and I need to shut down to safely remove the old USB drive and install the new one correct?  I can't (or shouldn't) pull the dead USB drive and put in the new one with Unraid still up.

 

Once I do shut down/install new USB drive/ reboot I would reassign all the drives appropriately (since it's still up I was able to screen shot the drive assignments and dump a diagnostic so I can be confident of assigning them correctly). After that do I go to tools 'new config' or do I just run a parity check (dual parity)?

 

Thanks for any help you can offer.

[tdallen]

First, unRAID keys are tied to the GUID of the flash drive.  You would need to migrate your licence from your old flash drive to the new one here.

 

That said, I don't understand how you've decided that you have a dead flash drive while unRAID is still running.  Typically you'd make that diagnosis after it crashes and/or fails to boot.

 

What you'd really like to do now is perform a clean shutdown and then migrate the configuration files to the new USB.  The configuration files on your old flash drive have the array configuration and would indicate a clean shutdown, allowing you to start up cleanly on the new USB.  But if your flash drive has really failed completely then shutdown can't write to it and presumably you can't recover any configuration files off it... 

 

What makes you think that the flash drive is completely dead?

 

In the event that the old flash drive is truly dead then I would:

• Get a new key file from the link above
  
• Stop all processes that might be writing to your server
  
• Attempt to perform a clean shutdown of the server
  
• Remove the flash drive from the server after it is shut down
  
• Check out that flash drive one last time, run chkdsk etc...
  
• Assuming it is dead and all configuration files have been lost boot using the new flash drive
  
• You won't have to do a new config because the new flash drive doesn't have any config at all - so assign all your data drives and parity drives
  
• If you truly believe you have clean data drives and good parity, check the box that says "Parity is already valid" and have unRAID start the array.  Then do a non-correcting parity check to be sure things look good.
  
• If you aren't sure you have a clean shutdown and good parity, then don't check that box and when you start the array unRAID will start a parity sync to build parity from scratch.  WARNING, your data will be unprotected until the parity sync completes.
  

 

I'd recommend before you take any permanent actions that you post back here more about how you diagnosed the dead flash drive, and if you want to migrate to the new flash drive then go ahead and get a replacement key.

[jasonfox]

Thanks for the reply tdallen.

 

CA alerted me that the flash drive was dead/inaccessible.  Trying to access the flash through smb or through the unraid main menu folder icon shows an empty drive. The machine is physically in the attic so I haven't pulled the actual drive to run a chkdsk yet. Unfortunately I hadn't set the backup flash feature in CA before this happened so I don't have a back up of the flash drives contents.      (That error in judgement has been corrected on my other two servers now though!    )

 

I had read the info at the link provided. My understanding is that I need to copy the previous key to the new drive and then once rebooted under the new drive go to tools>registration and select replace key to get the replacement.

 

Thanks for the bullet points on process going forward. The 'parity is valid' and 'non-correcting parity check after' is part I thought I had right in my head but needed confirmation on.

 

I am attaching the diagnostics file I pulled.  You can see in the syslog that the drive started failing around 4am Nov 21st and was caught by CA the following morning at about the same time.

 

hobbes-diagnostics-20161124-2142.zip