On Thu, Aug 8, 2013 at 4:06 PM, Matthew Seaman <[email protected]> wrote:
> On 08/08/2013 13:05, Mark Felder wrote: > > On Thu, Aug 8, 2013, at 6:59, Trond Endrestøl wrote: > >> > >> I'm just guessing, but I doubt a jail would be able to create new ZFS > >> filesystems outside its own structure, if at all able. A jail would > >> however be allowed to (un)mount already existing filesystems within > >> its own structure, i.e. Pool/test1. > >> > > > > When I first reviewed his post I clearly confused "mounting" with > > "creating a new zfs filesystem". Is that even supposed to be permitted > > in a jail? I almost feel a sysctl disabling that by default would be > > nice... DoS by zfs filesystem creation/deletion, anyone? > > There's a 'zfs jail' command and a 'jailed' property you can set on a > ZFS which I believes allow you to manage that ZFS from within the jail. > I think that extends to creating other ZFSes beneath that one (which > would inherit the 'jailed' property), BICBW. > > Mostly I find it easier to just manage the ZFSes from the host system > but then again, I'm not really making very extensive use of jails. > > Cheers, > > Matthew > > _______________________________________________ > [email protected] mailing list > http://lists.freebsd.org/mailman/listinfo/freebsd-stable > To unsubscribe, send any mail to "[email protected]" > Yes, it is easier to manage ZFS datasets from the host system but in this case we are assigning a different jail to each customer. That jail should be able to receive snapshots. It was working fine so far with 9.1. -- George Kontostanos --- http://www.aisecure.net _______________________________________________ [email protected] mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-stable To unsubscribe, send any mail to "[email protected]"
