Hi,

 

I have one data disk and one parity disk and my initial load of data ( from a mounted NTFS drive) was taking forever, so I stopped the array and mounted disk 1, and am now getting much better copy performance. My question is, do I have it right that once everything is copied I want to use the dreaded "Recover" button to build parity from disk1 with its new data?

 

Thanks,

 

TW

If you write a single byte to a drive with the array stopped, you will have to completely rebuild parity when you start the array.

 

The only situations where you save time by disabling parity, is when you need to load up multiple disks with data and the total data you are moving to unRAID is greater than the size of the parity disk.

Hi,

 

I have one data disk and one parity disk and my initial load of data ( from a mounted NTFS drive) was taking forever, so I stopped the array and mounted disk 1, and am now getting much better copy performance. My question is, do I have it right that once everything is copied I want to use the dreaded "Recover" button to build parity from disk1 with its new data?

 

Thanks,

 

TW

If you are currently running with no parity disk assigned, and you subsequently stop your array and assign one, unRAID will automatically calculate parity, regardless of use of the "Start" or "Restore" option...  They are nearly the same when initially assigning a parity disk.

 

They are VERY DIFFERENT once a parity disk is assigned and parity has been successfully calculated.  Pressing "Restore" at that point throws away the parity data, as if you had never calculated it, and throws away almost any chance of you recovering from a drive failure with it. 

 

Most of the time, 99.9% of the time, unless you are removing a drive from your array, and not replacing it a new one, or moving drives to a different and previously unused piece of hardware (different port) with you will never use the button labeled "Restore"

 

So, once your data is copied, just assign the parity drive, and "Start" the array.

I re-read what you typed...  depending on what you did, my first comment might be wrong.

 

If you stopped the array, then using your unix commands mounted the data disk to your own mountpoint, and then wrote to it with the array stopped, and then re-started the array... then YES, you WILL need to use the "Recover" button... or at the least, do a Manual parity check, and expect MANY parity errors detected.

 

If you un-assigned the parity drive, then started the array, then copied your data while the parity disk was un-assigned, you would have had the same speed advantage. Then, when you re-assigned the parity drive it would have done a full calculation.

 

Joe L.

I have already assigned a parity drive.

 

From your response I'm coming to the conclusion that hitting Recover to build parity from the new data is what I want to do, instead of starting and running parity check and seeing all the errors. I may not save time overall, but stopping the array as I did means I can complete the copy operation and disconnect the drive I'm copying from sooner and get it back into the desktop where I need it. The time spent calculating parity from scratch isn't a concern.

 

In the future I'll take your advice and wait to assign a parity drive.

 

Thanks,

 

TW

It worked!

 

I took your advice and unassigned the parity drive then disconnected it just in case. Mounted the ntfs drive and my drive 1 and performed the copy, then shutdown, plugged in the parity drive and unplugged the ntfs drive. Once I was booted I logged in and assigned the parity drive, then hit recover, then started, and it went right into a parity check. A couple hours later everything works!

 

Thanks,

 

TW