Hello,

 

I need some help from the community. I replaced a 1TB drive this morning in my server with a precleared 3TB drive (Drive 6). Since doing that I now see Drive 6 and 7 and unmountable and am getting thousands of Disk7 read errors in the syslog. Is it possible to replace Drive 6 with the original 1TB Drive?

 

So went and put the 1TB drive back in slot 6 but cannot mount the array as it now knows it should be looking for a 3TB drive 

Did you save the diagnostics before rebooting? If yes post them, if not post the current ones.

 

Were there writes to the array during the rebuild? (dockers/vms count if they are using the array)

Did you save the diagnostics before rebooting? If yes post them, if not post the current ones.

 

Were there writes to the array during the rebuild? (dockers/vms count if they are using the array)

 

did not save the diagnostics for the last reboot and putting the 1TB drive back in I have attached the latest Diagnostics and there shouldn't have been any writes to the array that I know of.

unraid-diagnostics-20170119-1012.zip

Disk7 has pending sectors and needs to be replaced:

 

 

Device Model:     WDC WD25EZRS-00J99B0
Serial Number:    WD-WMAWZ0041947

197 Current_Pending_Sector  0x0032   200   200   000    Old_age   Always       -       34

 

Now if there were no writes to the array, parity is valid with the old disk, so:

 

-do a new config

-select retain all assignments

-assign old disk6

-very important, check "parity is already valid" before starting the array.

-start array

-stop array

-assign new 3tb disk to disk7 slot

-start array to begin rebuild

 

If the disk is unmountable don't format, we'll deal with that later.

 

Keep old disk7 intact, you should be able to copy most data if the rebuild fails.

 

And since disk6 looks OK it was probably a connection issue and you should have checked the disks SMART and its connections before replacing it. Or asked us what to do.

 

Since you seemed unaware of the pending sectors on disk7 I assume you don't have Notifications setup either. You must let unRAID tell you when you have imminent problems before you get so many of them that you can't recover. Setup Notifications!

Disk7 has pending sectors and needs to be replaced:

 

 

Device Model:     WDC WD25EZRS-00J99B0
Serial Number:    WD-WMAWZ0041947

197 Current_Pending_Sector  0x0032   200   200   000    Old_age   Always       -       34

 

Now if there were no writes to the array, parity is valid with the old disk, so:

 

-do a new config

-select retain all assignments

-assign old disk6

-very important, check "parity is already valid" before starting the array.

-start array

-stop array

-assign new 3tb disk to disk7 slot

-start array to begin rebuild

 

If the disk is unmountable don't format, we'll deal with that later.

 

Keep old disk7 intact, you should be able to copy most data if the rebuild fails.

 

Thank you for the assistance, I took those exact steps and now have the 3TB drive in slot 7 and a Drive rebuild is taking place.

 

@ Trurl: I will setup notifications, however it is strange that the server never red balled the bad drive in Disk slot 7.

@ Trurl: I will setup notifications, however it is strange that the server never red balled the bad drive in Disk slot 7.

unRAID only disables a disk when a write to if fails. This is because the disk no longer has valid data on it. The failed write, however, is used to update parity, so that write is actually in the parity array, but not on the disk.

 

After a disk is disabled, unRAID will emulate the disabled disk by calculating its data from the parity array, and it will never use the actual physical disk again until it is rebuilt . Any time you try to read from the disk, unRAID will read all of the other disks plus parity and calculate the data for the read, and any time you try to write to the disk, unRAID will calculate and update parity so the written data can be later read by calculating it from the parity array.

 

So, after a disk write fails, the data on the physical disk is no longer valid, but the data in the array is and that valid data can be rebuilt to the same or a new disk.