On 11/18/10 08:58 PM, Frank Middleton wrote:
Got another slightly unhelpful message you might be able to explain:
# pkg image-update --be-name b134b
DOWNLOAD PKGS FILES XFER (MB)
Completed 1601/1601 33161/33161 666.0/666.0
pkg: Unable to clone the current boot environment.
# beadm list
BE Active Mountpoint Space Policy Created
-- ------ ---------- ----- ------ -------
b134 NR / 14.92G static 2010-11-18 19:14
Oddly, zfs list -t snapshot shows no snapshots of the root pool.
Here's the history:
Identical machines A and B both running some ancient version of
OpenSolaris. One was updated to snv126 and cloned to the other.
A was image-updated to snv134 and has been running forever. B
has multiple boot disks and was running both snv126 and sxce124.
It was this snv126 that failed to update to snv134b even though its
twin had updated to snv134 with no problem. I gave up on the
snv126->snv134b update, cloned the working snv134 after zpool
delete, rebooted, deleted all snapshots, and ran pkg image-update
as above.
If you have multiple BEs or ZFS datasets that share the same name that's
going to confuse the heck out of libbe, which in turn is going to create
problems for pkg(1).
I'd also note that split datasets are not supported currently. So you
can't have /opt on its own filesystem and expect things to work
correctly right now.
Beyond that, I can't help you here. This is a tangled mess that's
outside of my area of expertise.
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
