Background Info Running FreeBSD 8.2-STABLE 8x Hitachi 2TB 5K3000 connected via AOC-USAS2-L8i Drives given a GEOM label, then whole drive given to ZFS Story So I’ve had two drives in my RAIDZ2 array fail simultaneously, and after panicking for a while I managed to get my array functioning in a degraded state, backed up the most important data to another place and then shut it down. I have RMA’d the two drives, but won’t be able to get replacement drives for another 2-4 weeks. Now I’d like the use of my fileserver again whilst waiting for the RMA'd drives, but I’m extremely hesitant to run it without any redundancy whatsoever. As the 2Tb Hitachis are not in stock anywhere, I’ve come up with the following alternatives and would like your ideas on which would be the best thing to do: 1. Buy a new (4k sector) 2tb drive and use it to replace one of the failed drives. Potential performance issues as it’s a 4k drive mixed in with 512 sector drives. Inability to use the gnop trick to align the drive because that might reduce the available capacity of the drive below that of the other drives? Perhaps this could be counterbalanced by not assigning the replacement drive a GEOM label which would give me an extra 512 bytes of leeway? 2. Buy a 3tb Hitachi drive (512 byte sectors) and use that to replace one of the failed drives. Wasted $$$ Not in stock anywhere locally, so making option 1 work would be preferable. Any input you could provide would be much appreciated, or if there's a better 3rd alternative i'd love to hear it!