Hi, I am trying to build a Proxmox server install running ZFS using one disk initially, then add another disk later (why? - I will be breaking another mirror and using one of its disk to start the Proxmox install, then, once all good, will use the second disk to build the mirror (I don't have any spare disks of the correct size to start from scratch with 2 disks). To make sure all goes well, I am first testing with a smaller disk setup, using a relatively old simple Dell T110 II server, booting in BIOS mode. I installed Proxmox with one disk (it partitioned it with gpt and I used zfs raid 0), then after install, I shutdown added the second disk (as I would do when doing it for real). I then mirrored the partitions, copied data from the first 2 smaller partitions, and added the third partition to zfs, and the mirror syncing says it is complete (zfs mirror seemed very quick, seems to only mirror used data on disk unlike mdraid?) I then pulled the first disk to see if I could boot on the second mirror, but no such luck. I am not that familiar with gpt, but I installed parted and all partitions seem to have the same sizes and flags. I used dd to copy the two smaller partitions. I am not sure what is on these smaller partitions (with flags bios_grup and boot, esp). I thought I would be able to mount at least the second esp one as I thought that this was a fat 32 partition? Booting into an Ubuntu live disk, looking at the partitions, they are all labelled zfs members even though only sda3 and sdb3 show up in zpool status. Not sure why (and I could not figure out how to mount them). I suspect that on the sdb drive, something in one (or both) of those partitions needs to be changed so it will boot on the second drive? The summary question I guess is, how do I add a second disk to create a mirror that enables me to be allow it to boot from sda and sdb like I can with mdraid... having a mirror that can't boot when disk sda goes bad won't be that useful :-(, though at least I can get to the big zfs partition to get data off . Thanks for any help.