Raidz1 to Raidz2

[svargas]

I have to servers running 8.2 a production and backup server. The production box has a lot datasets with different permissions set and rsync tasks to replicate to the backup server, so data is safe and easy to copy back. I know I must delete the current VOL1 and create a new one to switch to raidz2. Is there away to put back the datasets and rsync tasks without recreating. If I rename the new volume the old one and reload the config would that put the datasets and rsync tasks in place?

Thanks for any help.....

 

[Sir.Robin]

The rsync tasks yes, but not the datasets.

If you have that many that its a lot of work, you could set up a temporary freenas and use zfs replication forth and back again.

That would recreate all your datasets if you replicate the entire volume.

Correct me if i'm wrong :)

 

[svargas]

So are you saying, that deleting the volume will not delete tasks.....

 

[Sir.Robin]

So are you saying, that deleting the volume will not delete tasks.....

Yes. The tasks are part of your config. So if you delete a volume, system settings, zfs replication and rsync tasks, and even defined shares will persist. Although cifs service should be disabled before deleting volumes or datasets that are shared.

 

[svargas]

Thanks for your reply. What about just adding another drive as a hot spare if that is an option in 8.2 which would provide two drives for failure.....

 

[Sir.Robin]

Raidz1 with hotspare will not protect you as good as raidz2 does.
Reason is because a failure with raidz1 will cripple your raid. If a second drive fails before or during rebuild, you have a lost array.
Not so with raidz2. You're still protected after first failure and will still be protected under rebuild.

Personally i would much rather have raidz2 than raidz1. To me it would be worth the hassle to set it up again.

Also, i'm not sure you can add a hotspare after the zpool is created?

I would set up another nas or another array. Replicate to the temp system/array. Rebuild original. Replicate back. That was simplified steps... you would need some configuring aswell and some downtime ;)

 

[svargas]

Thanks for your response again, you made good points that I did not think about. I don't have another server and drives to build another NAS. The reason I used raidz1 because I replicate to another server. I've had a couple drives go bad lately (old drives) replacing with new drives, as I hold my breath as it rebuilds.

 

[Sir.Robin]

You're welcome! Holding my breath is something i trie to avoid ;)

Also, ZFS checks on each read that the checksum vs the read data is intact. Othervise it knows a disk has problems.
This will secure your data. ZFS will repair files corrupted by a half-bad disk/bad sector as long as it has enogugh parity to calculate against.

This will not be true if you're running raidz1 and already has a failed drive.