I just completed a new 10TB Build using "UNFORTUNATELY" all WD EARS drives.

 

I am in the process of transferring 6.6TB to the new tower.  The problem I ran into is that all 10TB have been used up by the 6.6TB I transferred from my other server. (I should add that UNRAID is reporting the the correct drive sizes at ~2TB per drive).

 

I am running 4.7 with devices set at 4K aligned (without jumpers).  I didn't preclear "Probably a mistake" - and all drives are new out of the box.

 

I am transferring data using the rsync method and it has taken about ~75 Hours to complete.

 

I am assuming that I should preclear to check the drives and transfer the data again.  I would love to avoid this due to the amount of time it will require.

 

If anyone has suggestions Iwould love to hear them.

 

Thanks

JDM

You could pre-clear one drive at a time--so transfer off only one drive worth--or more everything off of one drive across the rest of your drives--and that backed up drive. Once (if) the drive passes, copy data off another drive, then pre-clear that one.

 

Rinse, repeat.

 

Tedious, yes, but better than doing a full transfer across the network again.

Thanks for the input. 

 

I just think its strange that the proper drive size is reported and the free space is not - Just to add - the reported free space was equal to the reported drive space before any file transfers were implemented.  Do you think there could be another issue or just an issue with one or more drives?

pre-clearing has absolutely nothing to do with free space.

 

If you insist on shot-gunning your fix, you'll get nowhere.

 

1.  post the output of the

df

command.

 

2. Attach a syslog to your next post.

 

3. Attach smartctl reports for each of your drives

 

4. Check the file-systems on each of your data disks.

 

Unless you know where all the space went, you'll never be confident of what is happening.

 

Odds are good you copied more than you think, or you have the min amount free settings wrong, or the split levels set wrong.

 

The only way to clear a drive now is to remove it from the array.  Doing that will cause you to lose parity protection.

 

Joe L.

Thanks Joe

 

I don't mind clearing the drives if that is the proper thing to do.  I don't have parity calculated yet and all the data that I need on this server exists on my primary server - I can transfer it again with no problem.

 

I copied the data via the instructions below - which I believe are yours and I don't think there is much chance that I copied more than I think because my primary server has over 3 TB free and it has a total of 10 TB storage as well

 

"Slightly harder, much faster, but still not difficult is to download and unzip the attached rsyncd.conf file to your flash drive on the new server.

Then, on the new server, log in and start the rsync daemon process by typing

rsync --daemon --config=/boot/rsyncd.conf

 

then on the old server with the files, you can invoke the rsync process to copy each user-share in turn as follows (my second server is named "tower2", you'll need to use the name of your second server in the following commands):

cd /mnt

rsync -rlpgoDvrH "user/Pictures" tower2::mnt/user/

rsync -rlpgoDvrH "user/Movies" tower2::mnt/user/

etc...

 

You'll need to leave the telnet session open while the copies are made, so disable any screen-shaver or sleep function on the PC with the telnet session, or, use the system console.

 

The rsync method is great in that it can be re-started if it is interrupted for any reason and checksums are automatically used.

"