Jonathan
2010-May-05 02:38 UTC
[zfs-discuss] Migrating ZFS/data pool to new pool on the same system
Can anyone confirm my action plan is the proper way to do this? The reason I''m doing this is I want to create 2xraidz2 pools instead of expanding my current 2xraidz1 pool. So I''ll create a 1xraidz2 vdev, migrate my current 2xraidz1 pool over, destroy that pool and then add it as a 1xraidz2 vdev to the new pool. I''m running b130, sharing both with CIFS and ISCSI (not comstar), multiple decedent file systems. Other than a couple VirtualBox machines that use the pool for storage (I''ll shut them down), nothing on the server should be messing with the pool. As I understand it the old way of doing iSCSI is going away so I should plan on Comstar. I''m also thinking I should just unshare the CIFS to prevent any of my computers from writing to it. So migrating from pool1 to pool2 0. Turn off AutoSnapshots 1. Create snapshot -> zfs snapshot -r pool1 at snap1 2. Send/Receive -> zfs send -R pool1 at snap1 | zfs receive -F -d test2 3. Unshare CIFS and remove iSCSI targets. For the iSCSI targets, seems like I can''t re-use them for Comstar and the reservations aren''t carried over for block devices? I may just destroy them before hand. Nothing important on them. 4. Create new snapshots -> zfs snapshot -r pool1 at snap2 5. Send incremental stream -> zfs send -Ri snap1 pool1 at snap2 | zfs receive -F -d test2 Repeat steps 4 and 5 as necessary. 6. Offline pool1... if I don''t plan on destroying it right away. Other than zfs list, is there anything I should check to make sure I received all the data to the new pool? -- This message posted from opensolaris.org
Apparently Analagous Threads
- After a 'virsh blockpull', 'virsh snapshot-list --tree' o/p does not reflect reality
- Re: Live snapshot merging (qemu 2.0)
- Managing Live Snapshots with Libvirt 1.0.1
- Re: virsh can't support VM offline blockcommit
- [help] Does "virsh blockpull" works on live virtual machine