XigmaNAS

Is there a method to replace a working online drive in a pool with one of higher capacity aside from remove->replace->resilver? It seems to me that there should be a method that involves reading ONLY from the drive being replaced, so as to not incur stress on all drives in the system each step of the way.

For reference, I have an 8x3TB raidz setup that I'm attempting to upgrade to 8x5TB. Originally I had planned on building a second system and zfs sending the data over, but I can no longer do that (lacking the hardware resources).

I'd rather not have all data read and re-read 8 times to finish this process (it seems it may take ~4 weeks to complete this way).

The only way I see, is to shutdown the server then try to dd each individual drive to a bigger one and then expand the filesystem, but I would not go that way.

this would transfer the array to the new drives, but i am not sure if you can then autoexpand the drives.

a better solution is to set up a new array and copy the data using zfs send. do you have 8 ports free? or a second controller?

You do have a backup right? Would it be easier to restore from backup?

dd does not sound like a good idea.

This is all valueless home data, so there's no full backup of the system.

I'm looking in to trying the zfs send method. Is it significantly faster than other transfer mechanisms (CIFS/NFS/FTP/etc)?