As I read Giovanni's mail, I don't think he's referring to the
"zone roots on zfs datasets" issues (as described in bug 6356600).
I think he's just asking whether a zfs dataset added to a zone's
environment via the "add fs" command in the zone's configuration
will also appear in that same zone in the new boot environment that
results from a lucreate. As far as I know, the answer is yes,
but someone more familiar with zones might have to answer
for sure. If the zone configuration is stored in a file in the root file
system, it would be copied to the new BE, and thus I would expect
file systems mounted in the zones in the PBE would also appear
in those same zones in the ABE, but I want someone with zones
knowledge to confirm this.
Antonello Cruz wrote:
> I don't think live_upgrade(5) is zfs friendly. I've tried to use
> live_upgrade(5) with my zones in a zfs filesystem and it didn't worked.
> You could try to detach the zone before the luupgrade and then attach
> it in the new BE, but I wouldn't hold my breath.
> Jeff Cheeney wrote:
>> Maybe someone on the install or zones discussion lists can help answer
>> this question.
>> Giovanni Schmid wrote:
>>> I have read different articles/docs/posts about solaris zones and
>>> liveupgrade issues until now; however, I have some doubts about the right
>>> way to deploy liveupgrade boot environments in the following case.
>>> I have a system with two disks configured as "mirrors" (that is, the same
>>> fdisk partition and VTOC).
>>> On the primary disk, I installed Solaris 10 8/07 with two sparse root
>>> zones, say Z1 and Z2. Just two file systems were settled on the primary
>>> disk: an UFS mounted on /, and a ZFS pool mirroring slice 4 of the two
>>> disks and mounted on /zfspool on the 1st disk. The UFS is intended to
>>> contain all but user's homes. These are served through a ZFS, namely
>>> /zfspool/users/home. Only zone Z2 inherits this ZFS, via the "add fs"
>>> All that premised, my questions are:
>>> What is the correct way of using Live Upgrade for this case ? Would
>>> something like:
>>> # lucreate -c bootenv1 -m /:c2d0s0:ufs -n bootenv2
>>> be sufficient ? That is, will Z2 in bootenv2 "see" /zfspool/users/home ?
>>> Any help is appreciated !
>>> This message posted from opensolaris.org
>>> storage-discuss mailing list
>>> [EMAIL PROTECTED]
>> zones-discuss mailing list
> install-discuss mailing list
> [EMAIL PROTECTED]
zones-discuss mailing list