Hi,
I'm replacing my current FreeNas setup with a new one, new hardware, new disks. I'll switch from a RAIZ-1 4-disk setup to a 2 disk mirror setup.
Once disk tests finishes, I'll have to migrate my config, data and jails.
I have one pool consisting of many hierarchical datasets.
As I wont have enough sata ports to connect all drives together, to migrate data I have two options: transfer from LAN or temporarily connect my new mirror pool in and USB enclosure (which I have). Which one works best?
As of now I have played a bit with snapshot replication. I have discovered that "recursive" replication do replicate sub-datasets, but it does not keep hierarchy on destination pool, all is laying flat. Have I missed a parameter?
What I understand is I have to:
- Create a new pool with the same name as the old one (or rename it afterwards with "zpool export" then "zpool import")
- Manually create the same dataset hierarchy on the new pool
- Create a one time replication task in the gui for each dataset (or I would prefer command line as it's much quicker and safer and allows for batch execution)
- Remove read only flag for each dataset
- Detach old pool from gui
- Connect old boot drive to the new system
- On the new system, import new pool
- Change network adapter name for each jail (?)
Will everything be OK?
I mean as I keep config I should keep users, groups, SMB and NFS shares, jails, ...?
Please correct me if I'm wrong somewhere or if there is a way to copy all at once to the new pool.
Julian
I'm replacing my current FreeNas setup with a new one, new hardware, new disks. I'll switch from a RAIZ-1 4-disk setup to a 2 disk mirror setup.
Once disk tests finishes, I'll have to migrate my config, data and jails.
I have one pool consisting of many hierarchical datasets.
As I wont have enough sata ports to connect all drives together, to migrate data I have two options: transfer from LAN or temporarily connect my new mirror pool in and USB enclosure (which I have). Which one works best?
As of now I have played a bit with snapshot replication. I have discovered that "recursive" replication do replicate sub-datasets, but it does not keep hierarchy on destination pool, all is laying flat. Have I missed a parameter?
What I understand is I have to:
- Create a new pool with the same name as the old one (or rename it afterwards with "zpool export" then "zpool import")
- Manually create the same dataset hierarchy on the new pool
- Create a one time replication task in the gui for each dataset (or I would prefer command line as it's much quicker and safer and allows for batch execution)
- Remove read only flag for each dataset
- Detach old pool from gui
- Connect old boot drive to the new system
- On the new system, import new pool
- Change network adapter name for each jail (?)
Will everything be OK?
I mean as I keep config I should keep users, groups, SMB and NFS shares, jails, ...?
Please correct me if I'm wrong somewhere or if there is a way to copy all at once to the new pool.
Julian
Last edited:
, don't know why it was not the case on local disk.
