Resizing file based extent
Posted: 19 Dec 2013 22:44
I'm in the process of lab testing a Nas4Free server that I'm about to implement at work for Clonezilla/Trilead backup/archival purposes.
My build is based on a single ZFS pool with a single lzjb compressed dataset. I'd like to present file based extents through iSCSI so I can extend/resize the "volumes" as needed. The FreeNAS WIKI states that this can be done by stopping the iSCSI service and running the "truncate -s +2T /dev/fu/bar" command on the extent. However, the "truncate" command doesn't appear to exist on Nas4Free (command not found). If truncate isn't the supported way of extending the extent, what's the supported way of resizing a file based extent?
FYI, at first I was using ZFS volumes instead of datasets because I could easily resize them. However, with ZFS volumes it kind of defeats the purpose of compression because you have to explicitly carve out volumes. Even if you're running 2.0x compression and the dataset/pool is only showing 50% utilization, the iSCSI client sees the volume as being full. That's why I think I need to use datasets and just "present" volumes of a certain size to the clients. That way I can over-present and advertise more space than is actually available so I can utilize the compression.
Am I totally on the wrong track? Could someone steer me in the right direction?
My build is based on a single ZFS pool with a single lzjb compressed dataset. I'd like to present file based extents through iSCSI so I can extend/resize the "volumes" as needed. The FreeNAS WIKI states that this can be done by stopping the iSCSI service and running the "truncate -s +2T /dev/fu/bar" command on the extent. However, the "truncate" command doesn't appear to exist on Nas4Free (command not found). If truncate isn't the supported way of extending the extent, what's the supported way of resizing a file based extent?
FYI, at first I was using ZFS volumes instead of datasets because I could easily resize them. However, with ZFS volumes it kind of defeats the purpose of compression because you have to explicitly carve out volumes. Even if you're running 2.0x compression and the dataset/pool is only showing 50% utilization, the iSCSI client sees the volume as being full. That's why I think I need to use datasets and just "present" volumes of a certain size to the clients. That way I can over-present and advertise more space than is actually available so I can utilize the compression.
Am I totally on the wrong track? Could someone steer me in the right direction?