I'm trying to get a feel on how to deal with a ZFS root filesystem when booted off an alternate medium. For UFS, this simply meant finding the correct device (slice on a disk) to mount and then mount it, assuming there wasn't some volume manager in the way.
For ZFS, this is a little more complex, especially if you don't want to modify the state of anything in the root pool. The root pool can be imported with zpool -R. However, simply importing the pool won't cause the root filesystem itself to get mounted. I assume this is because of the default canmount=noauto setting on it. If you try to simply zfs mount the root filesystem, you'll get something like this: cannot mount '/tmp/rpool': directory is not empty That's because the other stuff in the root pool *did* get mounted on the import, so root is unable to get mounted over it. So the question here is, what would be the proper way to mount the root filesystem (somewhere) without changing the state of anything in the pool ? The only way I seem to be able to get / mounted is to use zfs set mountpoint=<dir>, but this changes the mountpoint for / in the pool. My observation is that this doesn't really matter with root, but I'd rather not do this nonetheless. Furthermore, if I have a separate /var in the root pool that also needs to be mounted, doing a zfs set mountpoint for /var *does* cause a problem when booting back off of the root pool. It would seem that the answer would be the ability to run zfs mount with a -o mountpoint=<dir>, but this option doesn't seem to exist. Any thoughts on this ? Am I missing something here ? BTW, I'm currently trying this with the Solaris 10u6_04. It's possible that this may already be resolved in a later version (like OpenSolaris), but I need to be able to do this with Solaris 10. Thanks, John This message posted from opensolaris.org _______________________________________________ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
