I currently have a zpool with two 8Gbyte disks in it. I need to replace them with a single 56Gbyte disk. with veritas I would just add the disk in as a mirror and break off the other plex then destroy it. I see no way of being able to do this with zfs. Being able to migrate data without having to unmount and remount filesystems is very important to me. Can anyone say when such functionality will be implemented? This message posted from opensolaris.org
Mike DeMarco wrote:> I currently have a zpool with two 8Gbyte disks in it. I need to replace them with a single 56Gbyte disk. > > with veritas I would just add the disk in as a mirror and break off the other plex then destroy it. > > I see no way of being able to do this with zfs. > > Being able to migrate data without having to unmount and remount filesystems is very > important to me. > > Can anyone say when such functionality will be implemented? >If the original pool is a mirror, the it is trivial and has been a features since day one. zpool attach the new disk. zpool detach the old disks. If the original pool is not a mirror, then it can get more complicated, but depends on what you want it to look like in the long term... -- richard
> Mike DeMarco wrote: > > I currently have a zpool with two 8Gbyte disks in > it. I need to replace them with a single 56Gbyte > disk. > > > > with veritas I would just add the disk in as a > mirror and break off the other plex then destroy it. > > > > I see no way of being able to do this with zfs. > > > > Being able to migrate data without having to > unmount and remount filesystems is very > > important to me. > > > > Can anyone say when such functionality will be > implemented? > > > > If the original pool is a mirror, the it is trivial > and has been a > features since day one. zpool attach the new disk. > zpool detach the old disks. > > If the original pool is not a mirror, then it can get > more > complicated, but depends on what you want it to look > like in the long term... > -- richard > _______________________________________________ > zfs-discuss mailing list > zfs-discuss at opensolaris.org > http://mail.opensolaris.org/mailman/listinfo/zfs-discu > ssoriginal pool is not a mirror it consists of two disks concat. (SANS disks mirrored on SANS) Looks like attach will only allow attacment to a single device in a zpool and not to both devices. If I attach to one of the two devices it will grow the pool into the whole device and then I can not remove the other device. I find this to be a most frustrating problem with zfs and wonder if any work is being done to correct such a omission. This message posted from opensolaris.org
Mike DeMarco wrote:>> Mike DeMarco wrote: >> >>> I currently have a zpool with two 8Gbyte disks in >>> >> it. I need to replace them with a single 56Gbyte >> disk. >> >>> with veritas I would just add the disk in as a >>> >> mirror and break off the other plex then destroy it. >> >>> I see no way of being able to do this with zfs. >>> >>> Being able to migrate data without having to >>> >> unmount and remount filesystems is very >> >>> important to me. >>> >>> Can anyone say when such functionality will be >>> >> implemented? >> >>> >>> >> If the original pool is a mirror, the it is trivial >> and has been a >> features since day one. zpool attach the new disk. >> zpool detach the old disks. >> >> If the original pool is not a mirror, then it can get >> more >> complicated, but depends on what you want it to look >> like in the long term... >> -- richard >> _______________________________________________ >> zfs-discuss mailing list >> zfs-discuss at opensolaris.org >> http://mail.opensolaris.org/mailman/listinfo/zfs-discu >> ss >> > > original pool is not a mirror it consists of two disks concat. (SANS disks mirrored on SANS) > Looks like attach will only allow attacment to a single device in a zpool and not to both devices. If I attach to one of the two devices it will grow the pool into the whole device and then I can not remove the other device. I find this to be a most frustrating problem with zfs and wonder if any work is being done to correct such a omission. > >One solution is to split the new disk into two slices. Then attach each of the old disks to a different slice. This can be done with zero downtime. If you can handle a short downtime, then you can copy to a new, single slice file system. Today, ZFS treats the top-level vdevs as dynamic stripes, but there is no notion of a second-level dynamic stripe. On the one hand, this greatly simplifies data management. On the other hand, moving from two to one device exposes the top-level stripe. -- richard