How should I proceed with the backup?

[johan56]

Hi!

So I ran into a little problem a couple a weeks ago: https://forums.freenas.org/index.php?threads/running-hardware-raid-big-mistake.51885/

And now I'm about to back-up all data, and in the making also upgrade to FreeNAS Corral.

I have manage to get hold of 4 SATA II disk 2tb.

Im planning on attaching them to the onboard SATA controller, and make a ZFS pool of them.

Then run dd to clone the old disk, into the newly created ZFS pool(?)

Code:

> sudo dd if=/dev/disk0 of=/dev/disk2 bs=128m conv=noerror,sync

What do you guys think?

Any other precautions?

 

[m0nkey_]

You want to use zfs send and zfs recv to transfer your data.

 

[johan56]

Alright thanks.

Put thats seems to be when backing up off site, to another server. I would like to backup internally to newly added disks.

Or ami wrong?

 

[m0nkey_]

You're wrong.

Take a snapshot and send it.

# zfs snapshot tank/data@backup
# zfs send tank/data@backup | zfs recv backup/data

 

[Robert Trevellyan]

thats seems to be when backing up off site

zfs send and zfs receive work with input and output streams, which is an interoperability mechanism that many utilities in Unix-like operating systems implement. Combined with SSH, they can be used for offsite replication, which demonstrates the power of this mechanism.

 

[johan56]

Al

You're wrong.

Take a snapshot and send it.

# zfs snapshot tank/data@backup
# zfs send tank/data@backup | zfs recv backup/data

Alright! Thansk friendo!

 

[johan56]

Alright so I selected the "folder" under the volume, clicked "create snapshot", and this is what I got:

Diden´t have much success with

Code:

zfs snapshot raiden/raiden@backuppen

. I received

Code:

cannot open 'raiden/raiden': dataset does not exist

, but dosen´t it?

So next step is to plugin the spare disk, create a zfs pool named data. and run

Code:

zfs send raiden/raiden@backuppen | zfs recv backup/data

?

Thanks!

 

[m0nkey_]

zfs snapshot -r raiden@backuppen

If you want to capture all datasets, you only need to specify the pool name once and set the -r (recursive) flag.

 

[johan56]

Okey so I ran:

Code:

zfs send raiden@backuppenn | zfs recv ubuntuDiskar/data

and that went of instant! I thought it would take all night...

Any way, now it looks like this:

Is it safe for me to remove the "old" disks now?

 

[m0nkey_]

Is it safe for me to remove the "old" disks now?

No.

zfs send -R raiden@backuppenn | zfs recv ubuntuDiskar/data

You need to use -R recursive in your zfs send command.

 

[johan56]

It seems that everything went smoothly, except that once migrated to Corral, it was scrapped. And that I thought that I had encrypted disks, and I lost the passphrase and key. But it seems that all data is intact.

But I have another question regarding both the files that freeness created, and my plugins dataset.

How about the .system folder? Can I delete that?

And I used to have my jails and plugins in the dataset, on the volume I store my data. But I have read somewhere that you should have different volumes for that, is this correct? Can I just delete that dataset also?