So I was having an issue with replacing a drive in my array with a new larger drive, it caused a kernel panic and pre-clear failed.  I think it's an issue with ESX pass-through because I build a test bare metal box and pre-cleared and added the drive to the test array.

 

I then popped the disk out of the test array and was able to rebuild my old disk onto the new one on my prod unraid VM.  The rebuild had finished thsi morning and I went to run the diskspeed script to see how things had improved and I get this:

 

 

So I stopped the array and started in maintenance mode to run rieserfsck and it tells me I need to use --rebuild-sb.  I've also noticed that the drive is appearing an unformatted in the webui.

 

Any ideas?  I would post my syslog but it's 502KB and I'm not sure what part is relevant.

Just a thought - are you trying to pass though the drive, or the controller to which it is attached?

 

I am asking as when I tried to pass through disks rather than the controller to which they were attached it did not work correctly for drives larger than 2TB.

I'm passing through both of my LSI cards.

The array is still functioning with drive5 showing as unformatted so I went and bought another 3TB drive.  It's pre-clearing now and I'm hoping to replace disk5 with it and all the data will rebuild and be fine.  This time to do the drive swap/rebuild I'm going to do it on bare metal then once the rebuild is done go back to ESX.

Rebuilding the drive from parity will recreate exactly what is there now, so any file system errors will be the same. Unformatted is unraid's poorly phrased attempt at letting you know the file system can't be mounted, it doesn't really mean unformatted. Find a post by dgaschk and follow the directions in his sig line for a file system check.

see my sig