Ok here's what happened. I've built my system and I'm testing everything out using the Unraid basic so I have 3 400GB drives. Disk1 is parity, Disk 2 and 3 data. Disk 4 is in my case but obviously not used. I had created a folder called movies with on DVD rip in it. I also had a music folder with about 700 albums in it. All of this was on Disk 2. The shares were working fine and I could access all data. So far so good. Well I wanted to test a failure to see the recovery work. I then hit the power button for Disk 2. I then checked the //tower console and it showed an astrick by the DISK 2. I then pulled that drive and put the 4 th disk I had into that slot. It didn't seem to do anything so I started a parity check. It chugged away overnight. This morning it's reporting everything is fine but when I check the shares the movie and music folders are there but there isn't any of the subfolders or data. Can anyone explain what happened? Did I do something wrong? Thanks.

To replace a drive you must shut down the array and power off the server.  Your drive trays may support hot power up/down, but I don't think unRaid knows how to react when you do.

 

forcing a parity check would be asking it to build new parity with the contents of your current disks.  i'm hoping it asked to rebuild the defective drive.

 

in any case, try going to the shares page and press the re-scan button.  If you are lucky, your files will show back up in the user-share folders.    If the files are missing, and you don't see them on disk2 you can try this...

 

Shut down the array and power off the server.

Replace the now empty disk2 with the one you removed earlier.  It should still have your files, as long as you did not erase them somehow.

power back up the server.  It should notice the disk change and not start the array.

Now request that parity be rebuilt on your array.  it will chug along again and in a few hours your parity should be re-calculated and all your files you removed physically back again.

 

If you want to test the re-build, once parity is re-calculated, stop the array and then power down.

Remove disk2  and power up.  this will simulate running with a failed drive.

stop the array again and power down.

put the new disk (disk4) into the disk2 slot and power up.

unRaid should present you with the option of restoring to the new drive.

 

Joe L.

I'll try it. When this kind of situation happens are the options to do all these things presented in the //tower console or on the monitor that's attached to the unraid server. I didn't plan on having a monitor attached once it's setup.

Hey Joe, that is a nice explanation of steps to test out rebuilding a drive. I was wondering though, if you turn off an individual hard drive while the array is still running as str1der did, can you still then shut down the rest of the arrary, power off the server and begin the process you described? I was thinking when a hard drive does fail, it wont be during after a normal shutdown of a healthy array, but during everyday operation.

 

 

"Hotplug" is more-or-less supported in linux; it mainly depends on the controller used.  At present time however, hotplug is not supported by unRAID.  Plan is to support it in the future only with certain "qualified" hardware.

 

As for simulating drive failures... Probably easiest way is to Stop array, go to Devices page, and unassign on of your disks.  Go back to 'Main' and disk will appear Missing.  You can now Start array and observe data of Missing disk is still there (being provided by on-the-fly reconstruction).  Later you can Stop array, re-assign device, Start array, and observe reconstruction taking place.

 

As for real drive failures... the only drive failure which results in "disabling" a hard drive is an unrecoverable Write Error.